BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Computer Science and Engineering - ECPv6.13.0//NONSGML v1.0//EN
CALSCALE:GREGORIAN
METHOD:PUBLISH
X-WR-CALNAME:Computer Science and Engineering
X-ORIGINAL-URL:https://homecse.iitd.ac.in
X-WR-CALDESC:Events for Computer Science and Engineering
REFRESH-INTERVAL;VALUE=DURATION:PT1H
X-Robots-Tag:noindex
X-PUBLISHED-TTL:PT1H
BEGIN:VTIMEZONE
TZID:Asia/Kolkata
BEGIN:STANDARD
TZOFFSETFROM:+0530
TZOFFSETTO:+0530
TZNAME:IST
DTSTART:20250101T000000
END:STANDARD
END:VTIMEZONE
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251212T110000
DTEND;TZID=Asia/Kolkata:20251212T120000
DTSTAMP:20261010T130918
CREATED:20251207T103353Z
LAST-MODIFIED:20251207T103353Z
UID:2230-1765537200-1765540800@homecse.iitd.ac.in
SUMMARY:Relational Transformer: Toward Zero-Shot Foundation Models for Relational Data by Rishabh Ranjan
DESCRIPTION:Venue: Bharti501 \nAbstract: Pretrained transformers readily adapt to new sequence modeling tasks via zero-shot prompting\, but relational domains still lack architectures that transfer across datasets and tasks. The core challenge is the diversity of relational data\, with varying heterogeneous schemas\, graph structures and functional dependencies. In this talk\, I will present the Relational Transformer (RT) architecture\, which can be pretrained on diverse relational databases and directly applied to unseen datasets and tasks without task- or dataset-specific fine-tuning\, or retrieval of in-context examples. RT (i) tokenizes cells with table/column metadata\, (ii) is pretrained via masked token prediction\, and (iii) utilizes a novel Relational Attention mechanism over columns\, rows\, and primary-foreign key links. Pretrained on RelBench datasets spanning tasks such as churn and sales forecasting\, RT attains strong zero-shot performance\, averaging 93% of fully supervised AUROC on binary classification tasks with a single forward pass of a 22M parameter model\, as opposed to 84% for a 27B LLM. Fine-tuning yields state-of-the-art results with high sample efficiency. Our experiments show that RT’s zero-shot transfer harnesses task-table context\, relational attention patterns and schema semantics. Overall\, RT provides a practical path toward foundation models for relational data. https://arxiv.org/abs/2510.06377 \nBio: Rishabh Ranjan is a 3rd year PhD student at Stanford University co-advised by Jure Leskovec and Carlos Guestrin and supported by the Amazon Core AI Fellowship. His research is on building foundation models for relational data\, which includes databases\, tables\, time series and graphs. Before Stanford\, he has spent time at CMU and IIT Delhi\, where he was the President’s Gold Medalist for 2022. https://rishabh-ranjan.github.io
URL:https://homecse.iitd.ac.in/event/relational-transformer-toward-zero-shot-foundation-models-for-relational-data-by-rishabh-ranjan/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251211T110000
DTEND;TZID=Asia/Kolkata:20251211T170000
DTSTAMP:20261010T130918
CREATED:20251209T113155Z
LAST-MODIFIED:20251209T113155Z
UID:2239-1765450800-1765472400@homecse.iitd.ac.in
SUMMARY:Multiparty Session Types: Separation and Encodability Results by Prof. Nobuko Yoshida
DESCRIPTION:Venue: Bharti501 / Teams link will also be shared \nAbstract: Multiparty session types (MPST) are a type discipline for enforcing the structured\, deadlock-free communication of concurrent and message-passing programs. Traditional MPST have a limited form of choice in which alternative communication possibilities are offered by a single participant and selected by another. Mixed choice multiparty session types (MCMP) extend the choice construct to include both selections and offers in the same choice. This talk presents a mixed-choice synchronous multiparty session calculus and its typing system\, which guarantees communication safety and deadlock-freedom. We then talk of expressiveness of nine subcalculi of the MCMP-calculus by examining their encodability (there exists a good encoding from one to another) and separation (there exists no good encoding from one calculus to another). The highlight is that the binary (2-party) mixed sessions by Casal et al. (2022) are strictly less expressive than the MCMP-calculus. \n\nJoint work with Kirstin Peters appeared in LICS’24 (https://arxiv.org/abs/2405.08104) \nAbout the speaker. Nobuko Yoshida is Christopher Strachey Chair of Computer Science in University of Oxford. She is an EPSRC Established Career Fellow and an Honorary Fellow at Glasgow University. Last 10 years\, her main research interests are theories and applications of protocol specification and verification. She introduced multiparty session types [ POPL’08\, JACM ] which received the Most Influential POPL Paper Award in 2018 (judged by its influence over the last decade). This work enlarged the community and widened the scope of applications of session types\, e.g. runtime monitoring based on Scribble (co-developed with Red Hat) has been deployed to other projects such as cyberinfrastructure in the US Ocean Observatories Initiative (OOI); and widened the scope of her research areas. She received the Test-of-time-award from PPDP’24 and the best paper awards from CC’20\, COORDINATION’23 and DisCoTech’23. She received the third Suffrage Science Award for Mathematics and Computing from MRC for her STEM activity. She is an editor of ACM Transactions on Programming Languages and Systems\, Theoretical Computer Science\,  ACM Formal Aspects of Computing\, Mathematical Structures in Computer Science\, Journal of Logical Algebraic Methods in Programming\, and the chief editor of The Computer-aided Verification and Concurrency Column for EATCS Bulletin.
URL:https://homecse.iitd.ac.in/event/multiparty-session-types-separation-and-encodability-results-by-prof-nobuko-yoshida/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251209T140000
DTEND;TZID=Asia/Kolkata:20251209T150000
DTSTAMP:20261010T130918
CREATED:20251208T052758Z
LAST-MODIFIED:20251208T052758Z
UID:2233-1765288800-1765292400@homecse.iitd.ac.in
SUMMARY:Approximating Optimal Broadcast of Files in a Hose-Model Network
DESCRIPTION:Speaker:Sukriti Gupta (PhD student)\, CSE Dept.\, IIT Delhi\nAbstract -\nWe consider the problem of file sharing among peers who are connected to \na common core network through links of differing upload and download \ncapacities\, as is the case in networks provisioned according to the hose \nmodel. The file is assumed to be divided into equal-sized chunks\, and a \npeer can start sending a “chunk” of the file to another peer only after \nit has received the entire chunk. The objective is to share a chunk\, \ninitially residing on one of the peers\, with all other peers in the \nleast time possible. Peers can simultaneously send/receive parts of a \nchunk to/from multiple peers\, subject to the upload and download \ncapacity constraints.
URL:https://homecse.iitd.ac.in/event/approximating-optimal-broadcast-of-files-in-a-hose-model-network/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251201T160000
DTEND;TZID=Asia/Kolkata:20251201T170000
DTSTAMP:20261010T130918
CREATED:20251129T134307Z
LAST-MODIFIED:20251129T134307Z
UID:2218-1764604800-1764608400@homecse.iitd.ac.in
SUMMARY:Unifying Large Language Models and Knowledge Graphs for Faithful and Interpretable Reasoning by Gholamreza (Reza) Haffari
DESCRIPTION:Venue: SIT001 \nAbstract: Large Language Models (LLMs) demonstrate strong general reasoning ability\, yet still suffer from hallucination\, limited faithfulness\, and a lack of interpretability—especially in knowledge-intensive or domain-specific settings. Knowledge Graphs (KGs)\, on the other hand\, provide structured\, explicit\, and verifiable representations of facts\, but are incomplete and lack linguistic flexibility. This talk presents recent advances in unifying these two paradigms to achieve trustworthy and interpretable reasoning. In the first part\, I will introduce Reasoning on Graphs (RoG) and Graph-Constrained Reasoning (GCR)\, two frameworks that guide or constrain LLM reasoning using KG structure. RoG enables planning–retrieval–reasoning with faithful relation paths\, while GCR enforces KG-valid reasoning during decoding\, eliminating hallucinated reasoning paths and improving accuracy and interpretability. The second part of the talk presents GFM-RAG\, a graph foundation model trained on 60 diverse KGs with over 14 million triples for efficient\, multi-hop retrieval-augmented generation. GFM-RAG achieves state-of-the-art performance across multiple QA benchmarks and generalizes zero-shot to new datasets. Together\, these methods highlight a path toward unified\, scalable\, and reliable KG-LLM reasoning. \nBio:  Gholamreza (Reza) Haffari is a Professor in the Department of Data Science and Artificial Intelligence at Monash University\, Australia. He is a former ARC Future Fellow and previously served as Director of the Vision and Language Group. His research sits at the intersection of Natural Language Processing\, Deep Learning\, and Machine Learning\, with funding from ARC\, Google Research\, Amazon\, eBay\, Adobe\, and other industry partners. Reza also serves as the Chief AI Scientist at Openstream AI.
URL:https://homecse.iitd.ac.in/event/unifying-large-language-models-and-knowledge-graphs-for-faithful-and-interpretable-reasoning-by-gholamreza-reza-haffari/
LOCATION:SIT 001\, Amar Nath and Shashi Khosla School of Information Technology\, IIT Delhi\, Hauz Khas\, New Delhi 110016\, India\, Delhi\, Delhi\, 110016\, India
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251201T120000
DTEND;TZID=Asia/Kolkata:20251201T130000
DTSTAMP:20261010T130918
CREATED:20251129T134926Z
LAST-MODIFIED:20251129T134926Z
UID:2220-1764590400-1764594000@homecse.iitd.ac.in
SUMMARY:The End of "Seeing is Believing" - Securing Identity in the Age of Generative AI by Richa Singh
DESCRIPTION:Venue: SIT001 \nOnline Link: https://teams.microsoft.com/meet/4492945053702?p=Gz8szja9ebX7NoIitC \nAbstract: “In an era where AI can synthesize hyper-realistic faces and voices\, does the axiom ‘seeing is believing’ still hold water?” This question strikes at the very foundation of digital trust. Over a two-decade journey in biometrics\, contributing to large-scale systems like India’s Aadhaar and national security applications\, I have watched the challenges evolve from simple identification to complex verification. We have moved from solving “hard” recognition cases\, such as matching photos to forensic sketches or identifying individuals after plastic surgery and severe injury\, to confronting a far more insidious threat: the weaponization of generative AI. In this talk\, I will dissect the mechanics of this new adversarial landscape. I will present our work on detecting multimodal forgeries\, analyzing subtle visual artifacts\, temporal inconsistencies\, and acoustic anomalies in multilingual synthetic speech. Finally\, I will widen the lens to discuss the policy frameworks required to survive this shift—proposing a roadmap for digital identity that balances technical robustness with fairness\, bias mitigation\, and data sovereignty. \nBiography: Richa Singh is a Professor in the Department of Computer Science and Engineering at IIT Jodhpur. Her research spans responsible artificial intelligence\, machine learning\, pattern recognition\, biometrics\, and medical image analysis. She is a Fellow of the IEEE\, IAPR\, NASI\, and INAE\, and is an ACM Distinguished Member. Her honors include the NASSCOM AI Gamechangers Award and the Facebook Award for Ethics in AI. She is Founding Co-Editor-in-Chief of ACM AI Letters and Associate Editor-in-Chief of Pattern Recognition. She has served as organizing committee member of several conferences including PC Co-Chair of CVPR 2022and also served as Vice President (Publications) of the IEEE Biometrics Council.
URL:https://homecse.iitd.ac.in/event/the-end-of-seeing-is-believing-securing-identity-in-the-age-of-generative-ai-by-richa-singh/
LOCATION:SIT 001\, Amar Nath and Shashi Khosla School of Information Technology\, IIT Delhi\, Hauz Khas\, New Delhi 110016\, India\, Delhi\, Delhi\, 110016\, India
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251201T110000
DTEND;TZID=Asia/Kolkata:20251201T120000
DTSTAMP:20261010T130918
CREATED:20251124T162136Z
LAST-MODIFIED:20251124T162136Z
UID:2215-1764586800-1764590400@homecse.iitd.ac.in
SUMMARY:Building the Quantum Software Stack – and Verifying It by Prof. Ramanathan S. Thinniyam
DESCRIPTION:Venue: Bhart501\nAbstract: Quantum computing is approaching an inflection point. Global investment is scaling up\, hardware platforms are maturing – and the software stack is beginning to form. But what exactly is this stack? Who is building it\, and what remains to be done? In the first part of this talk\, I will give an overview of the emerging quantum software ecosystem: from quantum programming languages and compilers to simulators\, error mitigation\, and pulse-level control. I will highlight some of the current architectural directions in hardware (e.g.\, superconducting vs. neutral atom platforms) and outline the challenges of building reliable abstractions on top of noisy\, hardware-constrained systems.\n\nIn the second part\, I will shift focus to my own research: the use of formal methods—particularly automata-theoretic techniques—in reasoning about quantum circuits. I’ll present recent work on verifying properties of quantum circuits and how ideas from classical program analysis can be extended to this new domain. Throughout\, I’ll try to convey both the excitement and the difficulty of building a rigorous foundation for quantum software. \nBio: Ramanathan S. Thinniyam is an assistant professor in the Division of Computer Systems\, Department of Information Technology\, Uppsala University. Prior to joining Uppsala\, he was a postdoc at the Max Planck Institute for Software Systems\, Kaiserslautern\, and a visiting fellow at the Chennai Mathematical Institute. He obtained his Ph.D. from the Institute of Mathematical Sciences\, Chennai. He works on theoretical problems arising from verification in the classical settings\, with a recent focus on verification of quantum circuits. He is broadly interested in topics at the intersection of logic\, computation\, and mathematics.  
URL:https://homecse.iitd.ac.in/event/building-the-quantum-software-stack-and-verifying-it-by-prof-ramanathan-s-thinniyam/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251127T120000
DTEND;TZID=Asia/Kolkata:20251127T130000
DTSTAMP:20261010T130918
CREATED:20251124T161843Z
LAST-MODIFIED:20251124T161843Z
UID:2212-1764244800-1764248400@homecse.iitd.ac.in
SUMMARY:Constructing Long Paths in Graphs Streams by Chhaya Trehan
DESCRIPTION:Venue: Bharti501 \nAbstract: In the graph stream model of computation\, an algorithm processes the edges of an n-vertex input graph in one or more sequential passes while using a memory that is sublinear in the input size. The streaming model poses significant challenges for algorithmically constructing long paths. Many known algorithms that are tasked with extending an existing path as a sub-routine require an entire pass over the input to add a single additional edge. This raises a fundamental question: Are multiple passes inherently necessary to construct paths of non-trivial lengths\, or can a single pass suffice? To address this question\, we systematically study the Longest Path problem in the one-pass streaming model. \nIn this problem\, given a desired approximation factor α\, the objective is to compute a path of length at least lp(G)/α\, where lp(G) is the length of a longest path in the input graph G.We give algorithms as well as space lower bound results for both undirected and directed graphs. Besides the insertion-only model\, where the input stream solely consists of the edges of the input graph\, we also study the insertion-deletion model\, where previously inserted edges may be deleted again. Our results include: \n1. We show that for undirected graphs\, in both the insertion-only and the insertion- deletion streaming models\, there are semi-streaming algorithms\, i.e.\, algorithms that use space O(n poly log n)\, that compute a path of length at least d/3 with high probability\, where d is the average degree of the input graph. These algorithms can also yield an α-approximation to Longest Path using space \tilde O(n^2/α). \n\n2. Next\, we show that such a result cannot be achieved for directed graphs\, even in the insertion-only model. We show that computing a (n^{1−o(1)})-approximation to Longest Path in directed graphs in the insertion-only model requires space Ω(n^2). This result is in line with recent results that demonstrate that processing directed graphs is often significantly harder than undirected graphs in the streaming model. \n\n3. We further complement our results with two additional lower bounds. First\, we show that semi-streaming space is insufficient for small constant factor approximations to Longest Path for undirected graphs in the insertion-only model. Last\, in undirected graphs in the insertion-deletion model\, we show that computing an α-approximation requires space Ω(n^2/α^3). \n\nBio: Chhaya has a PhD in mathematics from the London School of Economics & Political Science. Prior to doing a PhD she has worked in tech Industry for 8 years in various capacities. She is currently doing a postdoc at the Indian Statistical Institute\, Kolkata
URL:https://homecse.iitd.ac.in/event/constructing-long-paths-in-graphs-streams-by-chhaya-trehan/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251117T150000
DTEND;TZID=Asia/Kolkata:20251117T160000
DTSTAMP:20261010T130918
CREATED:20251112T104537Z
LAST-MODIFIED:20251112T104537Z
UID:2187-1763391600-1763395200@homecse.iitd.ac.in
SUMMARY:Rank bounds and Polynomial Identity Testing by Prof. Akash
DESCRIPTION:Venue: Bharti501 \nAbstract: Polynomial Identity Testing (PIT) is the problem of checking whether a given algebraic circuit computes the zero polynomial. The PIT problem has a myriad of applications\, such as algorithms for the perfect matching problem\, primality testing\, and learning algorithms for sparse polynomials. While there are efficient randomized algorithms for PIT\, there is no deterministic poly-time algorithm for general circuits. Derandomizing PIT is a foundational problem in theoretical computer science\, as it is also intrinsically related to lower bounds for algebraic circuits and the VP vs. VNP problem. \nIn this talk\, we will discuss the PIT problem for depth-4 circuits. I’ll talk about recent progress on proving rank bounds for depth-4 identities\, and the first deterministic poly-time algorithm for depth-4 circuits with top fan-in 3 and constant bottom fan-in. We will discuss algebraic-geometric ideas such as the Stillman uniformity principle\, that lead to rank bounds and non-linear generalizations of classical results from combinatorial geometry.\n\nBio: Prof. Akash is an Assistant Professor in the Department of Computer Science at Rutgers University. I am part of the Theory group at Rutgers. His research interests are in Mathematics and Theoretical Computer Science\, in particular algebraic geometry\, computational complexity theory\, coding theory\, combinatorics and number theory. 
URL:https://homecse.iitd.ac.in/event/rank-bounds-and-polynomial-identity-testing-by-prof-akash/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251110T120000
DTEND;TZID=Asia/Kolkata:20251110T130000
DTSTAMP:20261010T130918
CREATED:20251104T161156Z
LAST-MODIFIED:20251104T161156Z
UID:2181-1762776000-1762779600@homecse.iitd.ac.in
SUMMARY:Non-Closure properties in algebraic complexity by Dr. Prateek Dwivedi
DESCRIPTION:Venue: Bharti501\n\nAbstract: A central question in algebraic complexity theory is understanding the behaviour of polynomial computation models under basic algebraic operations. While closure under addition and multiplication holds for most of the standard models like algebraic circuits\, closure under factorisation remains subtle. In this talk\, we will discuss a new result which proves that the well-studied model called read-once oblivious algebraic branching programs (roABPs) is not closed under factoring. This offers a contrasting perspective in light of the recent breakthrough work that proved a unified framework for analysing closure under factorisation. We will also discuss similar non-closure properties of roABP under other natural operations such as powering and symmetric composition. \nThis is based on joint work with Andrews\, Armand\, Hansen\, Limaye\, Srinivasan\, and Tavenas. \n[arxiv]
URL:https://homecse.iitd.ac.in/event/non-closure-properties-in-algebraic-complexity-by-dr-prateek-dwivedi/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251107T120000
DTEND;TZID=Asia/Kolkata:20251107T130000
DTSTAMP:20261010T130918
CREATED:20251011T113534Z
LAST-MODIFIED:20251011T113534Z
UID:2091-1762516800-1762520400@homecse.iitd.ac.in
SUMMARY:Can a Bucket of Water Translate? Exploring the Encoding–Decoding Ability of Randomly Initialized Neuro-Symbolic Transformers by Dr. Arghya Pal
DESCRIPTION:Venue: Bhart501 \nAbstract: There is a growing interest in developing artificial neural networks through the training of large models. But what lies hidden in an overparameterized neural network with random weights? If the distribution is properly scaled\, such a network contains a subnetwork that can perform well without ever modifying its weights. The number of possible subnetworks grows combinatorially with the size of the network\, and modern neural networks often contain millions or even billions of parameters. Thus\, we should expect that even a randomly weighted neural network contains a subnetwork capable of performing well on a given task. The core idea of this talk is to frame the subnetwork-finding problem as a Differentiable Integer Linear Programming (ILP) Solving problem. Unlike existing neuro-symbolic solvers\, this talk will introduce an algorithm that does not require a continuous relaxation of semantic constraints. Instead\, it allows for a direct\, more precise\, and efficient integration of neural representations into the ILP formulation. By the end of the talk\, we will see that the solver achieves superior performance compared to conventional ILP solvers\, neuro-symbolic black-box solvers\, and Transformer-based encoders. Furthermore\, a deeper analysis reveals that such a solver can significantly enhance the precision\, consistency\, and faithfulness of the generated explanations. This opens new opportunities for advancing neuro-symbolic architectures toward explainable and transparent deep learning in complex domains. \n  \nBio: Dr. Arghya Pal is a Lecturer at Monash University. He earned his Ph.D. from the Indian Institute of Technology Hyderabad\, followed by postdoctoral research appointments at Monash University and Harvard University.  His research focuses on generative models\, transfer learning\, causal inference\, learning under limited supervision\, and logical reasoning. He is a recipient of a gold medal for his Master’s degree\, the Intel Ph.D. Fellowship (2016–2020)\, honored as best researcher award twice – one during PhD and other as an Alumni from IIT Hyderabad. He has been honored with the Magna-cum-Laude award from Harvard MRI society and best researcher award from School of IT Monash University. \nDr. Pal has experience in teaching units like Modeling Discrete Optimization Problems\, Data Analytics\, Deep Learning\, Programming Paradigm\, Malicious Attack and Dark Sides of AI\, and Modeling Data Science Problems at Monash University. He assumed academic positions beyond teaching such as; session chair in IJCNN\, Senior Area Chair in AAAI\, Reviewer of prestigious venues such as TPAMI\, NeurIPS\, AAAI\, CVPR\, ICLR\, etc.
URL:https://homecse.iitd.ac.in/event/can-a-bucket-of-water-translate-exploring-the-encoding-decoding-ability-of-randomly-initialized-neuro-symbolic-transformers-by-dr-arghya-pal/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251103T120000
DTEND;TZID=Asia/Kolkata:20251103T130000
DTSTAMP:20261010T130918
CREATED:20251030T064546Z
LAST-MODIFIED:20251030T064546Z
UID:2178-1762171200-1762174800@homecse.iitd.ac.in
SUMMARY:Two-party cryptography beyond computational assumptions: Some old and new results by Dr. Akshay Bansal
DESCRIPTION:Venue: Bharti-425 \nAbstract: The impossibility of information-theoretic or unconditional security under classical communication is already established for many two-party cryptographic primitives\, including but not limited to coin flipping\, bit commitment\, and oblivious transfer. In this talk\, we first discuss the known limits of information-theoretic security using quantum communication and propose the novel framework of stochastic switching that uses stochastic semidefinite programming to develop simple protocols for various two-party tasks. We also briefly discuss the insufficiency of standalone security from the perspective of (in)composability of a weaker version of coin flipping. \nBio: Akshay Bansal is currently a Senior Scientist at a stealth venture based out of Bangalore. He recently completed his Ph.D. in Computer Science from Virginia Tech advised by Jamie Sikora\, with a research focus on quantum algorithms\, learning theory\, and convex optimization. He holds a Bachelor’s from IIT Kanpur and Master’s in Computer Science from ISI Kolkata. He has previously worked at the Centre for Quantum Technologies in Singapore\, Institute for Quantum Computing at University of Waterloo\, and held research position at IISc Bengaluru. His current work delves into the mathematical foundations of quantum cryptography\, classical and quantum machine learning\, and convex analysis.
URL:https://homecse.iitd.ac.in/event/two-party-cryptography-beyond-computational-assumptions-some-old-and-new-results-by-dr-akshay-bansal/
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251030T110000
DTEND;TZID=Asia/Kolkata:20251030T120000
DTSTAMP:20261010T130918
CREATED:20251028T055031Z
LAST-MODIFIED:20251028T055031Z
UID:2136-1761822000-1761825600@homecse.iitd.ac.in
SUMMARY:Towards Reliable LLM Reasoning: Coordinated Agents\, Variance-Aware Evaluation\, and Lean Inference by Prof. Akhil Arora
DESCRIPTION:Abstract: Large language models (LLMs) are increasingly deployed as reasoning engines\, yet their practical use remains constrained by three persistent challenges: achieving high-quality reasoning at low cost\, measuring performance reliably\, and ensuring efficient\, reproducible deployment. In this talk\, I will present a research agenda addressing these challenges through new methods\, benchmarks\, and systems for practical LLM reasoning. I begin with Next\, I turn to Finally\, I focus on Together\, these contributions chart a path toward LLM reasoning that is not only more powerful\, but also leaner\, more reliable\, and environmentally responsible.\n\n  \n\nBio: Akhil Arora is a Tenure-Track Assistant Professor of Computer Science at Aarhus University\, where he heads the CLAN for AI Research on Language and Networks (or “CLAN” for short). He is a fellow of the Copenhagen Center for Social Data Science (SODAS)\, an affiliate of the Pioneer Centre for AI and ELLIS\, and a formal collaborator of the Wikimedia Foundation\, the non-profit organization that manages Wikipedia and related projects. Akhil’s research lies broadly in human-centered AI with a focus on improving human knowledge-seeking\, bridging knowledge gaps\, and promoting knowledge equity on the Web. To this end\, he devises methods and tools blending techniques from NLP\, AI\, Graph ML\, and Computational social science. Recently\, his group has been devising robust\, trustworthy\, accessible\, and efficient LLM inference strategies.\nAkhil received his PhD in Computer Science from EPFL (2024) in Switzerland\, his MS from IIT Kanpur (2013)\, and his undergraduate degree from NCU Gurgaon (2010). In days of yore\, he spent close to five years in the industry working with the research labs of Xerox and American Express as a Research Scientist. His work on influence maximization has been recognized as the 8th most influential paper of SIGMOD 2017 by Paper Digest and received the 2018 ACM SIGMOD Most Reproducible Paper Award. He is a recipient of the prestigious EDIC Doctoral Fellowship\, an alumnus of the coveted Heidelberg Laureate Forum\, and a DAAD AINet fellow on human-centered AI. Akhil is a director of the P1-programs on
URL:https://homecse.iitd.ac.in/event/towards-reliable-llm-reasoning-coordinated-agents-variance-aware-evaluation-and-lean-inference-by-prof-akhil-arora/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251028T120000
DTEND;TZID=Asia/Kolkata:20251028T130000
DTSTAMP:20261010T130918
CREATED:20251016T070350Z
LAST-MODIFIED:20251024T203402Z
UID:2115-1761652800-1761656400@homecse.iitd.ac.in
SUMMARY:Adaptive Human-Robot Interaction: Human Inspired Handovers and Robotic Failure Explanations. (An Intersection of Robotics and Machine Learning) by Dr. Parag Khanna
DESCRIPTION:Venue: Bharti-501/MS Teams \nAbstract: \nAs robots become more advanced\, they are expected to be increasingly present among humans\, engaging frequently in physical and social interactions. Among these interactions\, handovers—the transfer of an object from one individual to another—play a vital role in daily life. This talk focuses on my research on enhancing human-robot interaction (HRI) by drawing inspiration from human-human handovers and utilizing handovers to resolve robotic failures by providing explanations for these failures as well as adapting these explanations based on human behavioral responses. \nFor physical interaction\, I present my work on formulating a weight-adaptive robot grip release strategy that determines when to release an object as a human recipient begins to take it and adapts to variations in object weight. I recorded and published datasets of human-human handovers to develop data-driven (LSTM\, VAE-LSTM based) grip release strategies\, which were experimentally validated in user studies. I also present how object weight affects human motion during handovers\, enabling robots to observe changes in human motion to estimate object weights and adapt their motions to convey weight information. Lastly\, I present my research on the use of non-touch modalities\, such as EEG brain signals and gaze tracking\, to discern human intentions during HRI\, differentiating between motions intended for handovers and those that are not. \nFor social interaction\, I explored how different levels of explanation content impact collaborative performance of human-robot teams and human satisfaction. I present my research on explanation variation strategies for repeated failures and adapting explanations by predicting user confusion. I further present a context-specific explanation generation system using behavior tree representation of collaborative tasks combined with Large Language Models (LLMs). This system was implemented as a failure communication module that enabled adapting explanation levels based on user queries and behavior\, effectively improving failure resolution rates for collaborative tasks\, as evaluated in user studies. \nBy this talk\, I aim to demonstrate how human-inspired approaches and machine learning can enhance both physical and social aspects of HRI\, and to outline my future research directions for adaptive HRI. \n  \nBiosketch:\nDr. Parag Khanna is a postdoctoral researcher at the Division of Robotics\, Perception\, and Learning (RPL) at KTH Royal Institute of Technology\, Sweden. As a collaborative roboticist\, his research combines insights from human behavior\, cognitive science\, and robotics to design intuitive and explainable robotic systems. His key research topics include physical and social human-robot interaction (HRI)\, analyzing and learning from human behavior\, data-driven and human-inspired robotic strategies\, and adaptive explanations for robotic failures. He is passionate about bringing robotics from the lab to everyday life through adaptive\, user-centered solutions that make robots more effective\, safer\, and easier to work with in real-world environments. \nHe received his PhD from KTH in 2025\, focusing on human-robot interaction—specifically\, developing adaptive techniques for seamless handovers between robots and humans. He holds dual M.Sc. degrees from the Erasmus Mundus European Masters in Advanced Robotics (EMARO+) program— from École Centrale de Nantes\, France\, and from the University of Genoa\, Italy (2019). He earned his B.Tech. in Mechanical Engineering from Visvesvaraya National Institute of Technology (VNIT)\, Nagpur\, India\, in 2017\, where his bachelor’s thesis on an autonomous snake robot reconfigurable into a quadcopter led to an Indian patent filed in 2017 (granted in 2025). \nFrom 2019 to 2021\, he worked as a research engineer at the French National Center for Scientific Research (CNRS) in Nantes\, France\, designing and controlling a bio-inspired tensegrity manipulator. \nHe has also organized workshops at the IEEE Humanoids conferences (2024–25) and the IEEE IROS 2025 conference\, and serves as a program chair for the HRI Pioneers Workshop at the HRI 2026 conference and as a Associate Editor for the IEEE/SICE System Integration (SII) 2026 conference. \nDr. Khanna’s accomplishments include the Honorable Mention for Best Short Contribution Paper Award and selection as a HRI Pioneer at the ACM/IEEE HRI conference 2025\, Best Poster Awards at KTH EECS Research and Impact Day 2025 and the Human Agent Interaction conference (HAI) 2023\, and representation of VNIT at the National Innovation Club member meeting at the Rashtrapati Bhavan in 2017.
URL:https://homecse.iitd.ac.in/event/adaptive-human-robot-interaction-human-inspired-handovers-and-robotic-failure-explanations-an-intersection-of-robotics-and-machine-learning-by-dr-parag-khanna/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251027T120000
DTEND;TZID=Asia/Kolkata:20251027T130000
DTSTAMP:20261010T130918
CREATED:20251023T103521Z
LAST-MODIFIED:20251024T205651Z
UID:2117-1761566400-1761570000@homecse.iitd.ac.in
SUMMARY:Sketching and Uncertainity: Through the Geometric Lens by Prof. Sujoy
DESCRIPTION:Venue: Bharti501\nAbstract: In many modern applications\, including machine learning\, robotics\, distributed systems\, and network design\, the input\, often represented as points in a finite metric space\, can bve massive in size. Efficient proceesing of such data requires compact representations that preserve the essential structural properties of the underlying space. Metric sketching provides a principled way to achieve this compression. Among the most fundamental sketching primitives are spanners and tree covers\, which capture distance relationships in a concise form. \nIn the first part of the talk\, I will discuss recent advances in geometric sketching. Traditional algorithmic models assume complete knowledge of the input in advance; however\, this assumption often fails in dynamic scenarios where inpiuts evolve over time. In such settings\, algorithms must adapt to changes while maintaining strong performance guarantees. \nIn the second part\, I will explore the dynamic aspects of metric sketching part\, related problems\, and highlight emerging directions that connect geometry with uncertainity.
URL:https://homecse.iitd.ac.in/event/sketching-and-uncertainity-through-the-gerometric-lens-by-prof-sujoy/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251014T120000
DTEND;TZID=Asia/Kolkata:20251014T130000
DTSTAMP:20261010T130918
CREATED:20251007T161139Z
LAST-MODIFIED:20251007T161139Z
UID:2079-1760443200-1760446800@homecse.iitd.ac.in
SUMMARY:Expanding the Frontiers of Computer Vision: From Robotics to Wildlife and Beyond by Ilan Shimshoni
DESCRIPTION:Abstract: In my talk I will describe in general how to cooperate with people from various fields of research in computer vision research projects and then describe three research projects that I was involved in in the last few years. \nIn the first project which is the field of archaeology we studied scarabs. Scarabs are seals whose origin is from ancient Egypt (2000 BC) but were also found in Israel. The dataset we obtained from an archaeologist consisted of pairs of a photograph and a drawing of a scarab made by an archaeological artist. We developed models for classifying the scarabs according to their etchings and according to the era when they were produced. The drawings are naturally of higher quality than the photographs. During training the model was fed with the photograph and the drawing and during inference only photographs were given as input\, since they are naturally more common. The algorithm also generated a drawing of the scarab. \nIn the second project\, which is in the field of agriculture\, a camera was placed above a drinking facility for sheep. The facility measures the amount of water the sheep drinks and its weight. A video of each sheep was recorded and its face\, back and legs were detected. The results of all these detections were fed into classifiers and the identity of the sheep was returned as a combination of the results from the single classifiers.  The process was basically automatic without human interaction. This algorithm can be used to monitor the condition of each sheep and report to the farmer if it seems that its medical condition has deteriorated. \nIn the last project\, which is in the field of ecology\, a colony of over a thousand terns on a small island was monitored. The terns fly from Europe to Africa and back and stay for some time on the island in Israel. The colony includes two types of terns: common terns and small terns. The whole island was scanned automatically using two PTZ cameras. Using Yolo the types of terns and whether they are brooding or not were classified. In a second stage the results improved since the actual size of the terns\, their motion pattern and their population statistics were taken into account. The results are very accurate and also include their geographic position on the island. This method can now be used to monitor the colony population over time. \nBio: Ilan Shimshoni has been working in the fields of computer vision\, computer graphics and machine learning for more than thirty years. He has been working on various problems in computer vision and applying them to applications in robotics and computer graphics. In recent years he has also been interested in addressing important problems in other fields which are challenging for researchers in my fields of research. He has been working for example on problems in medical rehabilitation\, geography\, agriculture\, and archaeology. One of my main fields of interest is developing algorithms in computer vision addressing challenges in the study of animals (wildlife\, pets\, and domestic animals). This include automatic detection of pain in cats and rabbits\, emotion in dogs\, and identification of individual sheep on a farm. In the realm of wildlife\, He has been working detecting flocks of birds from weather radars\, and counting terns of two types and identifying whether they are brooding or not. This helps ecologists estimate their number in one of the main places they stop in Israel while migrating from Europe to Africa.
URL:https://homecse.iitd.ac.in/event/expanding-the-frontiers-of-computer-vision-from-robotics-to-wildlife-and-beyond-by-ilan-shimshoni/
LOCATION:SIT 001\, Amar Nath and Shashi Khosla School of Information Technology\, IIT Delhi\, Hauz Khas\, New Delhi 110016\, India\, Delhi\, Delhi\, 110016\, India
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251013T120000
DTEND;TZID=Asia/Kolkata:20251013T130000
DTSTAMP:20261010T130918
CREATED:20251007T162719Z
LAST-MODIFIED:20251007T162719Z
UID:2088-1760356800-1760360400@homecse.iitd.ac.in
SUMMARY:Frontiers in Boolean Circuit Lower Bounds by Dr. Vaibhav Krishan
DESCRIPTION:Venue: Bharti501 \nAbstract:\nBoolean circuits provide a combinatorial representation of computation\, where the number of gates (the size) represents running time\, and the number of layers (the depth) capture parallel running time.\nThey form a framework for answering fundamental questions such as P vs NP: proving that some NP problem requires circuits of superpolynomial size would separate P from NP. \nWith limited progress on this question for general circuits\, early breakthroughs focused on restricted circuit classes.\nHåstad (STOC `86) proved that constant-depth circuits with AND\, OR\, and NOT gates require exponential size to compute the parity function (which determines whether the sum of inputs is even or odd).\nRazborov (Matematicheskie Zametki `87) and\, independently\, Smolensky (STOC `87)\, extended this to circuits augmented with parity or modular gates (for prime moduli)\, showing that such circuits require exponential size to compute the majority function (which determines whether the sum of inputs is at least half their number). \nFollowing these foundational results\, research has advanced along two principal directions\, though further progress has become increasingly challenging.\nIn this talk\, I will present some of my work contributing new advances at the frontier of both directions. \n______________________________________________________________________________________________________________________________________ \nPart I: Threshold Circuits. \nThe first part of the talk will focus on constant-depth threshold circuits\, circuits that can use majority gates\, or more generally\, threshold gates.\nThreshold gates can be seen as a simple abstraction for neurons\, and threshold circuits were among the earliest models studied to understand the computational power of neural networks.\nThese circuits are quite powerful; for instance\, they can efficiently implement integer arithmetic operations such as exponentiation and square root. \nIn joint work with Bajpai\, Kush\, Limaye\, and Srinivasan\, Algorithmica `21\, we study a generalization of threshold circuits\, called polynomial threshold circuits\, that use polynomial threshold gates.\nA polynomial threshold gate outputs a Boolean value based on the sign of a polynomial evaluated over the inputs.\nWe design an algorithm to count the number of assignments on which a polynomial threshold circuit outputs 1.\nFor any constant depth\, our algorithm runs faster-than-brute-force when the circuit size is slightly superlinear and the degree of each gate is bounded by a constant.\nPrior to our work\, no such algorithm was known even for a single polynomial threshold gate\, except in the special case of degree 2.\nFaster-than-brute-force algorithms are known to imply circuit lower bounds (Williams\, JACM `14)\, although the particular lower bounds implied by our work were already established by Kane\, Kabanets\, and Lu (STOC `17). \nOur work builds on a long line of research initiated by Impagliazzo\, Paturi\, and Saks (SIAM J. Comput. `97)\, who proved a tight lower bound for threshold circuits with a slightly superlinear number of wires computing the parity function.\nTheir core idea\, simplification of threshold circuits under random partial assignments\, has inspired a series of influential results\, leading to average-case lower bounds and satisfiability algorithms (Chen\, Santhanam\, and Srinivasan\, Theory Comput. `18)\, as well as pseudorandom generator constructions (Hatami\, Hoza\, Tal\, and Tell\, FOCS `22).\nEven seemingly small improvements to these results could lead to major breakthroughs in circuit complexity (Chen and Tell\, STOC `19)\, marking a central frontier for the community. \n______________________________________________________________________________________________________________________________________ \nPart II: Modular Circuits \nThe second part of my talk will focus on circuits with modular gates for general (not necessarily prime) moduli.\nHere\, in a joint work with S. Vishwanathan\, we develop an approach toward resolving a long-standing conjecture (Barrington\, JCSS `89)\, that constant-depth circuits with modular gates (the modulus does not grow with input size) require superpolynomial size to compute the majority function.\nThe classical lower-bound techniques of Håstad and Razborov-Smolensky fail to extend to this setting\, while the algorithms-to-lower bounds framework of Williams (JACM `14) does not apply to simple functions such as majority\, making new ideas necessary. \nVarious approaches to this conjecture have been proposed over the years.\nIn earlier work (Krishan\, CSR 2019)\, I showed that torus polynomials provide the most refined framework for tackling this problem.\nTorus polynomials were introduced by Bhrushundi\, Hosseini\, Lovett\, and Rao (ITCS `19) for proving lower bounds against such modular circuits.\nUsing torus polynomials\, we translate the lower bound conjecture to the task of finding feasible solutions to an infinite family of linear programs.\nThis reformulation allows for incremental progress\, by finding solutions for progressively larger sets from the family. \nWe find solutions for almost all of these programs\, leaving only a finite set unresolved.\nFinding feasible solutions for the remaining cases would lead to a resolution of the conjecture.\nTo make this task more tractable\, we show that the family of programs has far fewer degrees of freedom than initially expected\, and we describe a potential set of feasible solutions for some of the remaining cases.\nI will conclude the talk with open problems and directions for future progress. \n  \nBio: Vaibhav is a postdoctoral researcher at The Institute of Mathematical Sciences\, Chennai. He completed his Ph.D. at IIT Bombay under the supervision of Prof. Sundar Vishwanathan and Prof. Nutan Limaye. Before beginning his Ph.D.\, he worked as a quantitative researcher and a data scientist for four years.
URL:https://homecse.iitd.ac.in/event/frontiers-in-boolean-circuit-lower-bounds-by-dr-vaibhav-krishan/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251009T120000
DTEND;TZID=Asia/Kolkata:20251009T130000
DTSTAMP:20261010T130918
CREATED:20251007T162143Z
LAST-MODIFIED:20251007T162143Z
UID:2085-1760011200-1760014800@homecse.iitd.ac.in
SUMMARY:Incentives and Information in Algorithmics Economics by Dr. Divyarthi Mohan
DESCRIPTION:Venue: Bharti501/MS Teams \nAbstract: Digital markets and platforms have shaped the algorithmic landscape into a complex ecosystem of strategic\, self-interested entities. This has motivated the study and development of mechanisms or algorithms that are robust to strategic behaviour\, using tools from algorithms\, game theory and economics. Standard assumptions in mechanism design are too strong to capture the informational challenges present in many real scenarios\, from ad auctions where bidders’ values depend on competitors’ private market data\, to resource allocation where there is uncertainty about future demands. In this talk\, I will provide an overview of my recent work that tackles three important challenges—strategic behavior\, interdependence\, and online decision making—going beyond standard assumptions. In particular\, I will focus on my work establishing the first constant-approximation algorithms for prophet and secretary problems with interdependent values. \n  \nBio: Divyarthi Mohan is a postdoctoral researcher in the Faculty of Computing & Data Sciences at Boston University\, hosted by Kira Goldner. Her research broadly lies at the intersection of computer science and economics\, with a focus on algorithmic mechanisms design and the interplay of incentives and information. She obtained her PhD in Computer Science at Princeton University\, advised by Matt Weinberg\, and was previously a postdoctoral fellow at Tel Aviv University hosted by Michal Feldman. Her research has been recognized with the Simons-Berkeley Research Fellowship for Fall 2022\, the class of 2021 Siebel Scholarship\, and 2019 SEAS award for excellence at Princeton University\, and her work was invited to the Highlights Beyond EC 2024.
URL:https://homecse.iitd.ac.in/event/incentives-and-information-in-algorithmics-economics-by-dr-divyarthi-mohan/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20251007T120000
DTEND;TZID=Asia/Kolkata:20251007T130000
DTSTAMP:20261010T130918
CREATED:20251005T191731Z
LAST-MODIFIED:20251005T191731Z
UID:2082-1759838400-1759842000@homecse.iitd.ac.in
SUMMARY:Amnesiac Flooding and the curious case of a Unique Algorithm by Amitabh Trehan
DESCRIPTION:Venue: Bharti501 \nAbstract: In the field of distributed algorithm design\, it is often standard to abstract the network as an undirected graph with the nodes as vertices and connections as edges. About the simplest process one can imagine on a network/graph is flooding: A node is in possession of a message M which has to be  eventually sent to every node on the graph (this is called achieving broadcast)- the node sends M immediately to all its neighbours and they send to all their neighbours they did not just receive the message from and so on. Clearly\, this achieves broadcast. But\, how to achieve termination? i.e. the copies of the message should not circulate indefinitely. \n\n\nAt the advent of distributed computing\, more than 50 years ago\, a simple solution was devised – keep a copy of M\, and if M is received again\, simply discard this M. However\, this requires memory/state and a stack of earlier received messages. Surprisingly\, we discovered [PODC2019\,STACS2020\,DC2023] that state is unnecessary to achieve terminating broadcast – the same process without any state or memory beyond the immediate receipt (hence\, called Amnesiac Flooding (AF))\, due to some still slightly mysterious properties of simple undirected graphs\, terminates in asymptotically optimal time on every graph.  Intriguingly\, we have recently discovered [DISC2025] that AF is Unique! i.e. under certain reasonable conditions\, AF is the one and only algorithm that achieves terminating broadcast. Are there other examples of Unique algorithms in literature\, and is counting the number of algorithms for solving a problem a concept we can reasonably postulate? \n\n\nAF on Wikipedia: https://en.wikipedia.org/wiki/Amnesiac_flooding \n\nBio: Amitabh Trehan is an associate professor at the department of Computer Science\, Durham University\, where he heads the NESTiD (Network Engineering\, Science\, and Theory in Durham] research group. He did his PhD in Computer Science from the University of New Mexico\, USA\, following a M.Tech. in Computer Applications from the Indian Institute of Technology\, Delhi.
URL:https://homecse.iitd.ac.in/event/amnesiac-flooding-and-the-curious-case-of-a-unique-algorithm-by-amitabh-trehan/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250912T120000
DTEND;TZID=Asia/Kolkata:20250912T130000
DTSTAMP:20261010T130918
CREATED:20250912T045614Z
LAST-MODIFIED:20250912T045614Z
UID:1862-1757678400-1757682000@homecse.iitd.ac.in
SUMMARY:Chromatic number of randomly augmented graphs by Prof. Anand Srivastav\, Kiel University
DESCRIPTION:Abstract: An extension of the Erdős-Renyi random graph model Gn\,p is the model of perturbed graphs introduced by Bohman\, Frieze and Martin (Bohman\, Frieze\, \nMartin 2003). This is a special case of the randomly augmented graphs studied in this paper. An augmented graph is the union of a deterministic host graph \nand a random graph. Among the first problems in perturbed graphs has been the question how many random edges are needed to ensure Hamiltonicity of \nthe graph. This question was answered in the paper by Bohman\, Frieze and Martin. The host graph is often chosen to be a dense graph. In recent years \nseveral papers on combinatorial functions of perturbed graphs were published\, e.g. on the emergence of powers of Hamiltonian cycles (Dudek\, Reiher\, Ruciński\, \nSchacht 2020)\, the properties of Positional Games played on perturbed graphs (Clemens\, Hamann\, Mogge\, Parczyk\, 2020) and the emergence of multiple \ninvariants e.g. fixed clique size (Bohman\, Frieze\, Krivelevich\, Martin\, 2004). In this talk I will present our results on the chromatic number of randomly augmented \ngraphs. \n 
URL:https://homecse.iitd.ac.in/event/chromatic-number-of-randomly-augmented-graphs-by-prof-anand-srivastav-kiel-university/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250901T120000
DTEND;TZID=Asia/Kolkata:20250901T130000
DTSTAMP:20261010T130918
CREATED:20250806T062326Z
LAST-MODIFIED:20250809T094539Z
UID:1766-1756728000-1756731600@homecse.iitd.ac.in
SUMMARY:Giving Some Space Can Be Hard: Two New Models to Match Agents with Locations by Shivika Narang
DESCRIPTION:Abstract: There can be a multitude of reasons to match agents to specific locations in a given space. In this talk we cover two: distributing delivery orders and assigning shared hostel rooms. For both settings we shall try to find solutions that satisfy desirable properties and characterize instances for which they exist.\n \nWe first initiate the study of fair distribution of delivery tasks among a set of agents wherein delivery jobs are placed along the vertices of a graph. Our goal is to fairly distribute delivery costs (modeled as a submodular function) among a fixed set of agents while satisfying some desirable notions of economic efficiency. We characterize instances that admit fair and efficient solutions by exploiting underlying graph structures. Unfortunately\, finding these solutions proves to be NP-hard. We complement this by designing an XP algorithm (parameterized by the number of agents) that can find all fair and efficient solutions when they exist. We conclude this discussion by theoretically and experimentally analyzing the price of fairness.\n \nWe shall then introduce Leontief utilities to the problem of roommate matchings. We aim to find strategyproof mechanisms that give good bounds on agent welfare. We first find that no approximation to welfare can be achieved under strategyproof mechanisms for either Leontief or additive utilities. Even for binary additive utilities no maximum welfare mechanism can be strategyproof. In contrast\, we then –surprisingly– find that binary Leontief utilities enable us to find strategyproof mechanisms that maximize welfare.\n \nJoint work with Hadi Hosseini\, Sanjukta Roy and Tomasz Was.\n \nBio: Shivika Narang is a postdoctoral fellow at UNSW Sydney. Previously she was a postdoc at Simons Laufer Mathematical Sciences Institute\, Berkeley (SLMath) and completed her PhD from IISc Bengaluru. During her PhD\, she received the Tata Consultancy Services Research Fellowship. Her work is currently focused on finding fair and efficient solutions to societal problems. She largely works in computational social choice\, especially matching and allocation problems.
URL:https://homecse.iitd.ac.in/event/giving-some-space-can-be-hard-two-new-models-to-match-agents-with-locations-by-dr-shivika-narang/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250822T140000
DTEND;TZID=Asia/Kolkata:20250822T150000
DTSTAMP:20261010T130918
CREATED:20250827T082004Z
LAST-MODIFIED:20250827T082004Z
UID:1825-1755871200-1755874800@homecse.iitd.ac.in
SUMMARY:Optimal Capacity Modification for Stable Matchings with Ties by Dr. Keshav Ranjan
DESCRIPTION:Abstract: In this talk\, we consider the Hospitals/Residents (HR) problem in the presence of ties in preference lists of hospitals. Among the three notions of stability\, viz. weak\, strong\, and super stability\, we focus on strong stability. Strong stability is appealing both theoretically and practically; however\, its existence is not guaranteed. Our objective is to optimally increase hospitals’ quotas so that the resulting instance admits a strongly stable matching.\nSuch an augmentation is guaranteed to exist when resident preference lists are strict. We explore two natural optimization criteria:\n\n\n\nMINSUM: minimizing the total capacity increase across all hospitals and \nMINMAX: minimizing the maximum capacity increase for any hospital\n\nWe prove that the MINSUM problem admits a polynomial-time algorithm\, whereas the MINMAX problem is NP-hard. We prove an analogue of the Rural Hospitals theorem for the MINSUM problem. When each hospital incurs a cost for a unit increase in its quota\, the MINSUM problem becomes NP-hard\, even for 0/1 costs. In fact\, we show that the problem cannot be approximated to any multiplicative factor. We also present a polynomial-time algorithm for optimal MINSUM augmentation when a specified subset of edges is required to be included in the matching.\n\nThe talk is based on a recent work accepted at IJCAI 2025 and is a joint work with Meghana Nasre (IIT-M) and Prajakta Nimbhorkar (CMI). \nBio: Keshav Ranjan recently (July 2025) completed his Ph.D. from the Department of Computer Science and Engineering\, IIT Madras\, under the supervision of Dr. Meghana Nasre. His Doctoral thesis\, titled “Two-Sided Matchings: Lower Quotas\, Ties\, and Capacity Augmentation”\, focuses on the algorithmic aspects of two-sided matching problems under various constraints. Previously\, he held an M. Tech degree in Mathematics and Computing from the Department of Mathematics\, IIT Patna. His research interests lie in the broad area of Graph Algorithms\, with a particular focus on matching problems with preferences.
URL:https://homecse.iitd.ac.in/event/optimal-capacity-modification-for-stable-matchings-with-ties-by-dr-keshav-ranjan/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250814T120000
DTEND;TZID=Asia/Kolkata:20250814T130000
DTSTAMP:20261010T130918
CREATED:20250812T055310Z
LAST-MODIFIED:20250812T055310Z
UID:1773-1755172800-1755176400@homecse.iitd.ac.in
SUMMARY:A brief survey of quantum numerical algorithms by Pranav Singh
DESCRIPTION:Title: A brief survey of quantum numerical algorithms \nSpeaker: Prof. Pranav Singh \nDetails: August 14 (Thursday) | 12(noon)-1 PM | Bharti 501 \nAbstract:\nQuantum Numerical Algorithms (QNA) encompass a broad class of techniques including quantum numerical linear algebra (QNLA)\, quantum optimization\, quantum variational algorithms (QVA)\, quantum machine learning (QML)\, and Hamiltonian simulation (HS). These areas represent some of the most promising domains for realizing exponential quantum advantage and have seen rapid theoretical and algorithmic advances in recent years. In this talk\, I will provide a concise overview of these developments and highlight key challenges: both in terms of current quantum hardware limitations and the conceptual gap between classical numerical methods and emerging quantum paradigms. The goal is to offer both a technical snapshot of the field and a broader perspective on what makes quantum numerical thinking distinct and potentially transformative.
URL:https://homecse.iitd.ac.in/event/a-brief-survey-of-quantum-numerical-algorithms-by-pranav-singh/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250812T153000
DTEND;TZID=Asia/Kolkata:20250812T163000
DTSTAMP:20261010T130918
CREATED:20250809T092746Z
LAST-MODIFIED:20250809T092746Z
UID:1768-1755012600-1755016200@homecse.iitd.ac.in
SUMMARY:Enabling Energy-efficient AI Computing: Leveraging Application-specific Approximations by Akash Kumar
DESCRIPTION:Speaker: Akash Kumar (Ruhr University Bochum)\nDetails: Tue\, 12 Aug\, 3:30 PM\, SIT 001\n\nAbstract: The widespread adoption of Artificial intelligence and Machine Learning (AI/ML) models across various fields\, such as healthcare\, autonomous vehicles\, smart agriculture\, and industrial automation\, has led to a growing demand for efficient and scalable AI/ML solutions. However\, as AI/ML algorithms grow more complex\, their substantial memory requirements and high energy consumption pose significant challenges for deployment on resource-constrained embedded systems\, such as wearable health monitors and IoT devices. \nIn this talk\, I will first introduce the topic and outline the significance of cross-layer approximation framework\, emphasizing the necessity of a generic and scalable approach to designing approximate arithmetic operators. I will then talk about platform-specific optimizations for designing approximate operators optimized for FPGAs and end with how modern AI/ML-based DSE approaches can be used for approximate computer arithmetic. \nBiography: Akash Kumar received the joint Ph.D. degree in electrical engineering and embedded systems from the Eindhoven University of Technology\, Eindhoven\, The Netherlands\, and the National University of Singapore (NUS)\, Singapore\, in 2009. From 2009 to 2015\, he was with NUS. From October 2015 until March 2024\, he was a Professor with Technische Universität Dresden\, Dresden\, Germany\, where he was directing the Chair for Processor Design. Since April 2024\, he is directing the chair of Embedded Systems at Ruhr University Bochum\, Germany. His research interests include the design and analysis of low-power embedded multiprocessor systems and designing secure systems with emerging nano-technologies.
URL:https://homecse.iitd.ac.in/event/enabling-energy-efficient-ai-computing-leveraging-application-specific-approximations-by-akash-kumar/
LOCATION:SIT 001\, Amar Nath and Shashi Khosla School of Information Technology\, IIT Delhi\, Hauz Khas\, New Delhi 110016\, India\, Delhi\, Delhi\, 110016\, India
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250811T120000
DTEND;TZID=Asia/Kolkata:20250811T130000
DTSTAMP:20261010T130918
CREATED:20250809T094501Z
LAST-MODIFIED:20250809T094501Z
UID:1770-1754913600-1754917200@homecse.iitd.ac.in
SUMMARY:Explainable AI for Malware Analysis by Mohd Saqib
DESCRIPTION:Title: Explainable AI for Malware Analysis \n  \nAbstract: In recent years\, explainable artificial intelligence (XAI) has become a critical component of ensuring transparency and trust in machine learning systems\, particularly in high-stakes domains like cybersecurity. This talk will begin with a basic introduction to XAI\, highlighting its importance in understanding model decisions\, especially in the context of malware detection. I will then introduce GAGE (Genetic Algorithm-based Graph Explainer)\, a novel framework specifically designed for malware analysis. GAGE utilizes graph-based representations of malware features and applies a genetic algorithm to generate meaningful explanations for model predictions. This approach allows for both global and local interpretability of malware detection models\, making it easier for security professionals to understand how malware is identified and how detection decisions are made. The presentation will cover the theoretical foundations\, implementation details\, and experimental results of GAGE\, showcasing its potential to enhance trust and efficacy in automated malware detection systems. \n  \nBrief Bio: Dr. Mohd Saqib is a researcher and scholar specializing in Explainable AI (XAI)\, machine learning\, and cybersecurity. He completed his Ph.D. at McGill University\, where his research focused on developing interpretable models for malware analysis in collaboration with Defence Research and Development Canada (DRDC). Dr. Saqib also holds an M.Tech in Data Analytics from Indian Institute of Technology (ISM) Dhanbad. He has authored several Q1 journal papers\, including a comprehensive analysis of XAI for malware hunting published in ACM Computing Surveys (IF 23.8). Dr. Saqib has filed four U.S. patents in AI-related technologies during his collaborations with BlackBerry and Zayed University. In addition to his research\, Dr. Saqib has been a Teaching Assistant (TA) for cybersecurity\, hacking\, and AI courses at McGill University. He was honored with the Graduate Excellence Award at McGill and received the prestigious FRQNTscholarship. His expertise spans AI model explainability\, malware detection\, and the intersection of AI with critical infrastructure.
URL:https://homecse.iitd.ac.in/event/explainable-ai-for-malware-analysis-by-mohd-saqib/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250808T120000
DTEND;TZID=Asia/Kolkata:20250808T130000
DTSTAMP:20261010T130918
CREATED:20250804T083721Z
LAST-MODIFIED:20250804T083759Z
UID:1754-1754654400-1754658000@homecse.iitd.ac.in
SUMMARY:Multiturn Evals (and RL) for LLMs by Kartikeya Badola
DESCRIPTION:Title: Multiturn Evals (and RL) for LLMs \nDetails: 8th August\, 12 pm\, SIT001 \nAbstract: LLMs often fail at multi-step tasks requiring memory and strategic planning\, a gap not captured by traditional single-turn evals. To address this\, we’ve developed a suite of human and automated evals that stress test Gemini on these capabilities. This talk will cover the motivation and design behind these evals\, a discussion on latest results and will also touch upon some of the early promising experiments using multiturn RL methods to address some of these losses. \nBio: Kartikeya Badola is a Software Engineer at Google DeepMind in London\, where he works with the Gemini evals and Gemini thinking teams. Prior to this\, he was with Google Research in India\, working on multilingual semantic parsing. Kartikeya is a graduate of IIT Delhi\, where he worked with Prof. Mausam and Prof. Parag Singla on Distantly Supervised Relation Extraction. \n 
URL:https://homecse.iitd.ac.in/event/multiturn-evals-and-rl-for-llms-by-kartikeya-badola/
LOCATION:SIT 001\, Amar Nath and Shashi Khosla School of Information Technology\, IIT Delhi\, Hauz Khas\, New Delhi 110016\, India\, Delhi\, Delhi\, 110016\, India
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250806T120000
DTEND;TZID=Asia/Kolkata:20250806T130000
DTSTAMP:20261010T130918
CREATED:20250730T061826Z
LAST-MODIFIED:20250730T165814Z
UID:1727-1754481600-1754485200@homecse.iitd.ac.in
SUMMARY:Rank Aggregation and Fairness by Diptarka Chakraborty
DESCRIPTION:Abstract: Aggregating multiple input rankings over a set of candidates to generate a consensus ranking is one of the fundamental ranking problems\, having many applications in social choice theory\, hiring\, college admission\, web search\, and databases. However\, the optimal consensus ranking might be biased against any individual candidate or candidates belonging to certain marginalized communities or groups. This has motivated studies of the rank aggregation problem from the fairness perspective. While finding a consensus ranking\, the additional objective is to ensure fair representation of each group in the top positions of the final aggregated ranking. In this talk\, we will discuss various algorithms to find such a fair ranking approximately.\n\nSpeaker: Diptarka Chakraborty is an Assistant Professor at the National University of Singapore. He did his Ph.D. at the Indian Institute of Technology\, Kanpur. Before joining NUS\, he spent two years at Charles University\, Prague\, and then almost a year at Weizmann Institute of Science\, Israel\, as a post-doctoral fellow. His research interest mostly lies in theoretical computer science\, more specifically\, algorithms on large data sets\, approximation algorithms\, sublinear algorithms\, string matching algorithms\, and graph algorithms. He is a recipient of the best paper award at FOCS 2018 and the Google South & Southeast Asia Research Award 2022.
URL:https://homecse.iitd.ac.in/event/rank-aggregation-and-fairness/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250728T113000
DTEND;TZID=Asia/Kolkata:20250728T123000
DTSTAMP:20261010T130918
CREATED:20250721T150910Z
LAST-MODIFIED:20250728T034019Z
UID:1709-1753702200-1753705800@homecse.iitd.ac.in
SUMMARY:Dot-Product Proofs by Prahladh Harsha
DESCRIPTION:Title: Dot-Product Proofs\n\nSpeaker: Prahladh Harsha (TIFR Mumbai):  Details: July 28 (Monday) | 11:30 AM | Bharti 404 \n\n\nAbstract: A dot-product proof is a simple probabilistic proof system in which the verifier decides whether to accept an input vector based on a single linear combination of the entries of the input and a proof vector. In this talk\, I will present constructions of linear-size dot-product proofs for circuit satisfiability and discuss two kinds of applications: basing the exponential-time hardness of approximating MAX-LIN (maximal number of linear equations that can be simultaneously satisfied) on the standard exponential-time hypothesis\, and minimizing the verification complexity of cryptographic proof systems.\n\n[Joint work with Nir Bitansky\, Yuval Ishai\, Ron Rothblum\, and David Wu] \nBio: Prahladh Harsha is a Professor at the School of Technology and Computer Science at the Tata Institute of Fundamental Research (TIFR)\, Mumbai\, India. He obtained his BTech degree in Computer Science and Engineering from IIT Madras in 1998 and his PhD in Computer Science from the Massachusetts Institute of Technology (MIT) in 2004. He has worked at Microsoft Research\, TTI Chicago\, and has been at TIFR since 2010. \nPrahladh’s research interests are in the area of theoretical computer science\, with special emphasis on computational complexity theory and algebraic coding theory. He is best known for his work in the area of probabilistically checkable proofs. Prahladh Harsha is a winner of the NASI Young Scientist Award for Mathematics and the Swarnajayanti Fellowship (Govt. of India). \n\n\nProf. Harsha has served on the editorial boards of SIAM Journal on Computing and Algorithmica. He is currently the Editor-in-Chief of the ACM Transactions on Computation Theory and a Fellow of the Indian Academy of Sciences.
URL:https://homecse.iitd.ac.in/event/dot-product-proofs-by-prahladh-harsha/
LOCATION:Bharti 404\, IIT Delhi\, Delhi\, New Delhi\, 110016\, India
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250725T120000
DTEND;TZID=Asia/Kolkata:20250725T130000
DTSTAMP:20261010T130918
CREATED:20250724T051140Z
LAST-MODIFIED:20250724T051140Z
UID:1721-1753444800-1753448400@homecse.iitd.ac.in
SUMMARY:Thinking Geometrically in the World of Data-driven Models by Niloy J. Mitra
DESCRIPTION:Title: Thinking Geometrically in the World of Data-driven Models \nAbstract:\nAs learning models grow in scale and capacity\, a natural question arises: is it still relevant to think geometrically\, or should we seek more data? In this talk\, I will argue that geometry remains essential\, not just for interpretability\, but for enabling control\, structure\, and generalization. I will illustrate this through three examples. First\, how to distil image diffusion features to enrich geometric models by combining projection operation with diffusion features. Second\, in video generation\, 3D awareness proves critical for spatial control\, object permanence\, and task awareness in multimodal LLMs. Finally\, I will present a neural surface representation that exposes direct access to first and second fundamental forms\, enabling a new approach to defining Laplace operators on neural surfaces. Together\, these examples highlight that geometry continues to offer powerful tools\, perhaps more so now than ever. For more details\, please visit https://geometry.cs.ucl.ac.uk/. \nBio:\nNiloy J. Mitra leads the Smart Geometry Processing group in the Department of Computer Science at University College London and the Adobe Research London Lab. He received his PhD from Stanford University under the guidance of Leonidas Guibas. He was an assistant professor at IIT Delhi from 2007-2009. His research focuses on developing machine learning frameworks for generative models for high-quality geometric and appearance content for CG applications. He was awarded the Eurographics Outstanding Technical Contributions Award in 2019\, the British Computer Society Roger Needham Award in 2015\, and the ACM SIGGRAPH Significant New Researcher Award in 2013. He was elected as a fellow of Eurographics in 2021 and served as the Technical Papers Chair for SIGGRAPH in 2022. His work has also earned him a place in the SIGGRAPH Academy in 2023. Besides research\, Niloy is an active DIYer and loves reading\, cricket\, and cooking. For more\, please visit https://geometry.cs.ucl.ac.uk
URL:https://homecse.iitd.ac.in/event/thinking-geometrically-in-the-world-of-data-driven-models-by-niloy-j-mitra/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250609T120000
DTEND;TZID=Asia/Kolkata:20250609T130000
DTSTAMP:20261010T130918
CREATED:20250604T065636Z
LAST-MODIFIED:20250604T065657Z
UID:1639-1749470400-1749474000@homecse.iitd.ac.in
SUMMARY:Designing advanced cryptographic primitives in distributed settings
DESCRIPTION:Speaker: Dr. Anshu Yadav\, Postdoctoral Researcher\, Institute of Science and Technology\, Austria \nAbstract: In today’s world\, the rapid advancement of technology has led to the generation of vast amounts of sensitive data\, which must be accessed in a secure and controlled manner to facilitate research across various domains. Often this data\, associated with a single logical entity\, is generated in a distributed manner\, yet must be protected with the same level of security as if it were produced by a single source. Furthermore\, distributing authority among multiple entities is essential to avoid a single point of security failure. These are natural\, yet complex challenges in modern cryptography. My research focuses on exploring how advanced cryptographic primitives can provide effective solutions to such problems. In this talk\, I will begin with a brief overview of my research interests and profile. I will then focus on the themes discussed above. In particular\, I will briefly talk about a result on multi-input attribute based encryption which is a generalization of attribute-based encryption (ABE) – a novel encryption paradigm enabling expressive access control on encrypted data. In ABE\, a message m is encrypted under an attribute x\, and decryption keys are associated with a policy 𝑓. Decryption is possible if and only if 𝑓(x)=1\, unlike traditional public key encryption scheme where a single key can decrypt all ciphertexts. In the multi-input setting\, data is generated by k non-interacting parties\, with each party contributing an input (x_i\, m_i)\, so that x=(x_1\,…\,x_k) and m=(m_1\,…\,m_k). The function 𝑓 is now a k-ary predicate. The goal is for each party to independently encrypt their data as ct_1\,…\,ct_k\, and for a decryption algorithm with key sk_f to recover (m_1\,…\,m_k) if and only if 𝑓(x_1\,…\,x_k)=1. We formally defined the notion of multi-input ABE (k-ABE) and presented constructions for different k under different cryptographic hardness assumptions. I will describe the key challenges in designing cryptographic schemes in the multi-input setting and how our work addresses these challenges. If time permits\, I will also briefly talk about my work in threshold cryptography – a very useful and active field of cryptography with advanced practical applications in distributed environments (e.g.\, block chains\, distributed key generation\, etc.). In threshold cryptography\, a privileged operations—such as ‘signing’ in digital signature scheme or ‘decryption’ in an encryption scheme—is distributed among n parties\, ensuring that at least a threshold t of them are required to perform the operation. In this area\, my research has mostly focussed on threshold signatures\, where we improve the security of post-quantum threshold signature scheme. Finally\, I will conclude the talk with a discussion of my future research directions\, including open problems in the areas discussed above.
URL:https://homecse.iitd.ac.in/event/designing-advanced-cryptographic-primitives-in-distributed-settings/
LOCATION:Bharti 501\, IIT Campus\, Hauz Khas\, New Delhi
CATEGORIES:Seminars
END:VEVENT
BEGIN:VEVENT
DTSTART;TZID=Asia/Kolkata:20250605T100000
DTEND;TZID=Asia/Kolkata:20250605T110000
DTSTAMP:20261010T130918
CREATED:20250602T072215Z
LAST-MODIFIED:20250602T094001Z
UID:1633-1749117600-1749121200@homecse.iitd.ac.in
SUMMARY:Multimodal Learning in 3D environments: Perception\, and Simulation
DESCRIPTION:Speaker: Dr. Arun Balajee Vasudevan is currently a Research Scientist at Amazon \nAbstract: \nAutonomous robots have several potential applications such as virtual assistants\, VR/AR\, gaming\, self-driving technologies\, city planning and others. To achieve autonomy\, a robot needs to see and hear the environment\, before it can converse or navigate favorably to perform a human-desired task. Precisely\, scene understanding begins with building 3D geometry\, decoding semantics\, understanding surround objects/humans\, and planning and actions. My research talk addresses these fundamental challenges independently under two broad themes: understanding geometry and multimodal perception & data-driven simulation and navigation. \nUnder geometry and perception\, I introduce the usage of several multimodalities such as the user’s gaze\, visual sensors such as cameras or range sensors (e.g. Kinect)\, and speech/language instructions from human referrals for robot perception tasks. Further\, I delve deep into the investigation of the audio sensing modality for the task using binaural sound microphones. Secondly\, regarding the geometry\, my talk addresses one of my ongoing works about the construction of digital twins of the real world with 4D reconstruction of dynamic scenes from ground visuals of a robot. \nUnder the theme of Data-driven simulation and robot navigation. Following perception\, robots must navigate and take meaningful actions in the world. This involves broadly two aspects: wayfinding and motion planning. Earlier works address wayfinding based on directional instructions\, overlooking human aspects. I briefly talk about a new paradigm that integrates principles from cognitive science with learning-based methods to tackle the challenge of language-based wayfinding for robots in real-world outdoor environments. The second aspect is motion planning for which I propose the learning of driver behavior models for MPC-based planners to build data-driven simulators. \nLong term\, I envision to bridge the above two themes to build multimodal digital twins simulators of the real-world. This potentially helps in training and testing of planners\, VR/AR setups\, gaming\, and others. Lastly\, I also cover my future research plan in the talk. \nShort Bio: \nArun Balajee Vasudevan is currently a Research Scientist at Amazon. Previously\, he was a postdoctoral researcher at Carnegie Mellon University under Prof. Deva Ramanan till March 2025. His core research interest is in Computer Vision and Multimodal Learning. He has works in multimodal (vision\, language and sounds) perception and navigation\, 3D/4D reconstruction\, Motion Planning and improving Foundational models. He published papers predominantly in vision and machine learning conferences/journals such as CVPR\, ECCV\, ICML\, IJCV\, TPAMI\, and others. He defended his PhD under Prof. Luc Van Gool at ETH Zurich. He received his MSc in Computer Science from EPFL in 2016 and an undergraduate degree in Electrical Engineering from the Indian Institute of Technology Jodhpur in the year 2014.
URL:https://homecse.iitd.ac.in/event/multimodal-learning-in-3d-environments-perception-and-simulation/
CATEGORIES:Seminars
END:VEVENT
END:VCALENDAR