Mechanism Design and Deep Learning
Designing Optimal Auctions through Neural Networks
Fair Mechanism Design for Agriculture through Neural Networks
Building Fairness into AI: The Role of Mechanism Design
Introduction to Cooperative Game Theory
Cooperative Game Theory and Influence Maximization
Cooperative Game Theory and Explainable AI
A globally integrated supply or service industry concentrates on making supply chains efficient and competitive and slogan is competition is supply chain vs supply chain. Companies pursued strategies such as outsourcing, offshoring, and lean manufacturing to cut costs, retain market position, or gain competitive advantage.
The current day supply chain networks are subjected to disruptions from unknown and or known sources affecting the performance, profits, and their survival yet times. New Technologies: Mobile Internet, IOT, social media, big data analytics, Cloud computing, 3D printing, Algorithmic Governance, Blockchains, AI, LLM, etc. are changing the industry structure. Supply chains try to be risk resistant but over the last three years powerful external forces driving global impacts — the pandemic, military conflicts, a heightened cyber risk environment, trade wars, decades-high inflation, and the effects of accelerating climate change. The pandemic has shifted attention of the supply chains to Food, Pharma, Electronic, Healthcare, Online retail, etc. Several industries including Auto, Travel, Hospitality, Hotels, Restaurants went through a dip. Most countries have increased their defence spending. Unemployment and inflation are on the rise.
The network governance model is generally hierarchical with all the data and decision making are with one company. Currently Supply Chain Transparency; Resilience, Supply Chain Finance; Smart Contracts and Network Governance are getting attention. Also, Big data analytics using AI and machine learning to automate the managerial decision making is gaining attention.
In this lecture, our aim is to provide an ecosystem-based design of green resilient supply chain networks. The relevance of this lecture is high in the current situation. We develop the GRIP (Governance, Risk, Innovation, Performance) framework for the supply chain networks. The new performance measures include traceability, the risk assessment and mitigation framework. The modern supply chain needs modern tools to help companies assess, monitor, and proactively act upon risks. Finally, we present the governance mechanism for the entire supply chain using multi-sourcing, control towers and online hierarchical control methods. This area is highly attractive for researchers as well start-ups.
N. Viswanadham is Honorary Scientist in the Computer Science and Automation at the Indian Institute of Science. From 1967-1998, he was faculty at the Indian Institute of Science (IISc). Professor Viswanadham was Professor and Executive Director for The Center of Excellence Global logistics and manufacturing strategies in the Indian school Of Business, during 2006-2011. He was Deputy Executive Director of The Logistics Institute- Asia Pacific and Professor in Department of Mechanical and Production Engineering at the National University of Singapore during 1998-2005.
He is the recipient of the 1996 IISc Alumni award for excellence in research. He was conferred the Distinguished Alumni Award in 2009 by IISc. He was awarded the 2012 Prof S K Mitra Memorial Award by the INAE. He is a Fellow of the IEEE, a Fellow of Indian National Science Academy, Indian Academy of Sciences, Indian National Academy of Engineering, and The World Academy of Sciences.
Professor Viswanadham has made significant contributions to the areas of manufacturing, logistics and global supply chain networks. He is the author of Four Textbooks, Nine Edited Volumes, over two hundred forty journal and top tier conference papers.
His current research efforts are in the areas of Design of Future supply chain networks and Competitive Business Models including Platforms and Cooperatives. He also studies the application of the new technologies to healthcare, agriculture, and smart villages.
Proteins are made up of amino acids. Paralogous proteins are diverged from a common ancestor but have different functionalities. For a pair of paralogous proteins, we ask the question which key amino acids account for the differences in their functionalities. Towards this, we first propose a linear classifier with a 20 dimensional feature vector, based on the composition of each protein. First, we demonstrate that just the amino acid compositions, not even their positions in protein sequences, is adequate to distinguish 15 paralogous pairs of proteins from each other. Next, we use a recently proposed Shapley Value based Error Apportioning, SVEA, algorithm to identify a subset of key amino acids, Amino acid Feature Subset, AFS, for this classification task. The SVEA algorithm considers a classification game, a cooperative game, with features as players and identifies the key subset of features in a dataset based on the Shapley value of this classification game. This AFS (Amino acid Feature Subset) are the key amino acids that distinguish a pair of paralogous proteins. We validate the ability of the AFS amino acids to discriminate paralogous proteins by analyzing multiple sequence alignments of protein families and/or by providing supporting evidence from literature. We also pair-wise classify sub-families of a protein superfamily and highlight common amino acids identified in the AFS for two paralogous proteins within a common sub-family.
Based on an ongoing work with Pranav Machigal, Bussi Rakesh and Petety V. Balaji.
N. Hemachandra is a Professor of Industrial Engineering and Operations Research as well as a faculty member of Center for Machine Intelligence and Data Science at IIT Bombay. He held visiting positions at IISc and IIM Bangalore. His current academic interests are broadly in Decision Sciences, with a focus on sequential decision making models, including data-driven ones like Reinforcement Learning, Bandit problems, etc. His other academic interests also include various resource allocation problems arising in supply and value chains, communication networks, and logistics.
The Multi-Armed Bandit (MAB) problem is a classical model used to capture decision-making in uncertain environments. This problem involves an agent faced with k choices, referred to as the arms of the bandit. At each time step, the agent selects an arm to pull, receiving a reward specific to that arm. To maximize its total reward over time, the agent grapples with the explore vs. exploit dilemma: it can either choose to explore—pull an arm for new knowledge—or exploit—pull an arm based on existing knowledge, such as one that has yielded high rewards thus far. This model finds application in a diverse set of real-world scenarios, such as online advertising, crowd sourcing, and medical trials. This has led to the study of several variants of the classical MAB model, each with different modeling assumptions. One such novel variant is the Improving MAB model.
In the Improving Multi-Armed Bandit problem, the reward obtained from an arm increases with the number of its pulls. This model provides an elegant abstraction for many real-world problems in domains such as education and employment, where decisions about the distribution of opportunities can affect the future capabilities of communities and the disparity between them. A decision-maker in such settings must consider the impact of their decisions on future rewards in addition to the standard objective of maximizing cumulative reward.
In this talk, we will briefly explore the history of the MAB model and then delve into some of our recent results for the Improving MAB problem.
Based on a joint work with Vineet Nair, Ganesh Ghalme, and Arindam Khan.
Vishakha recently completed her PhD from the Department of Computer Science and Automation (CSA) at the Indian Institute of Science (IISc), where she was jointly advised by Prof. Y. Narahari and Prof. Arindam Khan. Her research focused on fairness and causality in online decision-making problems, including Multi-Armed Bandits and Markov Decision Processes. Her doctoral work was supported by a Google PhD Fellowship and a CII-SERB PM Fellowship for Doctoral Research. Vishakha previously completed her MTech (Research) from CSA, IISc, under the guidance of Prof. Narahari.
Abstract: Regret minimization is a pre-eminent objective in the study of decision making under uncertainty. Indeed, regret is a central notion in multi-armed bandits, reinforcement learning, game theory, decision theory, and causal inference. In this talk, I will present our recent work that extends the formulation of regret with a welfarist perspective.
In particular, we quantify the performance of a decision maker by applying a fundamental welfare function–namely the Nash social welfare (NSW)–and study Nash regret, defined as the difference between the (a priori unknown) optimum and the decision maker’s performance. Since NSW is known to satisfy fairness axioms, our approach complements the utilitarian considerations of average (cumulative) regret, wherein the algorithm is evaluated via the arithmetic mean of its expected rewards. I will present our recent works that obtains essentially tight Nash regret guarantees for stochastic multi-armed bandits (MAB) as well as linear bandits.
Including joint works with Arindam Khan, Arnab Maiti, and Ayush Sawarni, and Soumyabrata Pal.
Siddharth Barman is an Associate Professor at the Department of Computer Science and Automation at the Indian Institution of Science Bengaluru. He is the recipient of the ACM India Early Career Researcher (ECR) Award, 2023 and a Ramanujan Fellowship. Prior to joining IISc, he was a post-doctoral researcher at Caltech. He obtained his Ph.D. in Computer Science at the University of Wisconsin-Madison. His research interests lie at the interface of computer science and microeconomics, with a current focus on algorithmic aspects of fair division.
We develop a general framework for selecting a small pool of candidate solutions to maximize the chances that one will be optimal for a combinatorial optimization problem, under a linear and additive random payoff function. We formulate this problem using a two-stage distributionally robust model, with a mixed 0-1semidefinite program. This approach allows us to exploit the “diversification” effect inherent in the problem, to address how different candidate solutions can be selected to improve the chances that one will attain a high ex-post payoff. More interestingly, using this distributionally robust optimization approach, our model recovers the “evil twin” strategy, well-known in the field of football pool betting, under appropriate settings.
We also address the computational challenges of scaling up our approach to construct a moderate number of candidate solutions to increase the chances of finding one that performs well. To this end, we develop a sequential optimization approach based on a compact semidefinite programming reformulation of the problem. Extensive numerical results show the superiority of our approach over existing methods.
To demonstrate the efficacy of this approach, we use this approach to pick winners for popular lottery games. We demonstrate its advantage over random quick pick in the game of lotto.
This talk is based on recent joint works with Liu Changchun (SIA-NUS Digital Aviation Corp lab) and Liu Ju (NUS Chongqing Research Institute).
Chung Piaw Teo is Provost’s Chair Professor in NUS Business School and concurrently the Executive Director of the Institute of Operations Research and Analytics (IORA) in the National University of Singapore. He is Co-Director of the SIA-NUS Digital Aviation Corp Lab and Director of the Center on Modern Logistics in Chongqing-NUS Research Institute. With a focus on optimization and supply chain management, Professor Teo is trying to bridge the gap between theoretical research and practical applications of OR and Analytics in business and engineering.
He was a fellow in the Singapore-MIT Alliance Program, an Eschbach Scholar at Northwestern University (US), a Professor at Sungkyunkwan Graduate School of Business (Korea), and a Distinguished Visiting Professor at YuanZe University (Taiwan). He is the department editor for MS (Optimization) and a former area editor for OR (Operations and Supply Chains). He was elected Fellow of INFORMS and appointed Changjiang Scholar (USTC) in 2019 and awarded a Public Administration Medal (Silver) by the Singapore Government in 2023. He has also served on several international committees such as the Chair of the Nicholson Paper Competition (INFORMS, US), member of the LANCHESTER and IMPACT Prize Committee (INFORMS, US), Fudan Prize Committee on Outstanding Contribution to Management (China), and recently chaired the EIC search committee for Operations Research, an INFORMS journal
In mechanism design with transfers (such as auctions), though stochastic mechanisms are feasible in any optimization program of the designer, a deterministic mechanism often emerges as optimal in many settings. One of the reasons is that the set of incentive-constrained mechanisms form a convex set and its extreme points are deterministic. I will review some recent results explaining the (sub)-optimality of deterministic mechanisms.
I work on mechanism design, and have broad interests in all topics in microeconomic theory. Some of my current research is on multidimensional mechanism design, mechanism design for regulation problems, matching, and social choice theory. I did my PhD from the University of Wisconsin, Madison in 2004, and I have been at the Indian Statistical Institute since 2006.
More details can be found here:
https://www.isid.ac.in/~dmishra/cv.pdf
Natural disasters are common in India and annually affect perhaps the largest number of people in a country like ours with far-reaching consequences in terms of loss of infrastructure and resources such as agricultural lands, housing, roads, etc. One of the main concerns we have is how quickly can we reach the affected area and disseminate help and assistance in terms of essential resources. However, disasters often come unannounced and catch us by surprise. What is available is often much less than what we need. How do we distribute these scarce resources when there are so many contenders? In this talk, I will present a non-cooperative game model that would address this problem to some extent by producing pure strategy optimal solutions. Several examples from cyclones, floods, and COVID situations will be given to illustrate how these solutions can be implemented in real disaster scenarios.
Debasish Ghose did his BSc (Engg) from NIT Rourkela in 1982, ME, and PhD from IISc in 1984 and 1990. He is currently a professor at the Robert Bosch Center for Cyber-physical Systems the Department of Aerospace Engineering at the Indian Institute of Science. He has also held various visiting positions, such as at the University of California at Los Angeles. He is a fellow of INAE, NASI, and INSA and an associate fellow of AIAA. His research interests are in the areas of guidance of autonomous vehicles, robotics, game theory, and AI/ML applications. He is and has been, on the editorial boards of many IEEE Transactions and journals and is currently with the IEEE Transactions on Control of Network Systems.
Panini, who lived around the 4th century B.C., developed a technique for noun compounding using rules. This kind of noun compounding can also be found in other languages such as German. We have used this technique to generate ontologies and perform searches on large text corpora. In the recent past, we have also witnessed considerable progress in neural networks and the integration of these networks with ontologies and other knowledge structures. In this talk, I will present our work on natural language processing of large text corpora, and several machine learning techniques used for bio-medical imaging. Particular topics include: Improving measurement techniques, which generate terabytes of image data, used in cell biology; Using a Root&Rule-based technique based on Panini’s grammar for generating ontologies; Applying category theory for knowledge representation; Analyzing Trojan detecting algorithms in neural networks; Evaluating speech emotion recognition systems; and Discussing predictive coding for integrating neural networks with knowledge networks.
– Ram D. Sriram, Chief, Software and Systems Division, National Institute of Standards and Technology, Gaithersburg, MD, 20899
Ram D. Sriram is currently the chief of the Software and Systems Division, Information Technology Laboratory, at the National Institute of Standards and Technology. Before joining the Software and Systems Division, Sriram was the leader of the Design and Process group in the Manufacturing Systems Integration Division, Manufacturing Engineering Laboratory.
Before joining NIST, he was on the engineering faculty (1986-1994) at the Massachusetts Institute of Technology (MIT). Sriram has co-authored or authored more than 300 publications, including several books on artificial intelligence. Sriram was a founding co-editor of the International Journal for AI in Engineering. Sriram received several awards, including four-lifetime achievement awards. His most recent awards are: IIT Madras Distinguished Alumni Award (2021), the IEEE Reliability Society’s Lifetime Achievement Award (2023), and the 2024 Product Lifecycle Management Pioneers Award from IFIP TC5/WG5.1 for his groundbreaking work on computers and information modeling for design and manufacturing. Sriram is a Fellow of AAIA, ACM, AIBME, ASME, AAAS, IEEE, IET, INCOSE, SMA, SME, and Washington Academy of Sciences, a Senior Member (life) AAAI, and an Honorary Member of IISE. Sriram has a B.Tech. from IIT, Madras, India, and an M.S. and a Ph.D. from Carnegie Mellon University, Pittsburgh, USA.
In the last two decades, we have witnessed two exponential laws in technology (informatics and genomics) both reach critical scale. When technologies have improved by doubling some 32 times, we have the phenomenon of exponential scale up which begins to resemble singularities. We call this the democratization of technology because at this stage becomes affordable and accessible. The emergence of deep learning, generative AI, next-generation genome sequencing, CRISPR gene editing, LLMs, and quantum computing are appearing now in rapid succession. This is now a great time for multidisciplinary, team-driven innovation to build spectacular solutions as recombinations of these advanced tools. And this is at a time when the globe faces great challenges of climate change, health, environment, and food security. The speaker will relate his experiences and journeys with intelligent and living machines and their future potential.
Vijay Chandru (PhD MIT), is an academic, an entrepreneur, and an advisor in translational research in Bangalore, India. He has taught as a tenured professor at Purdue and at the Indian Institute of Science (IISc). He is a fellow of the National Academies of Science (IASc 1996) and Engineering (INAE 2010) in India and the American Association for the Advancement of Science (AAAS 2023). A Technology Pioneer of the World Economic Forum and recipient of the President’s Medal of INFORMS in 2006, Chandru was given the Hari Om Trust award by UGC for contributions to science and society in 2003.
Vijay is a visiting scientist at Harvard School of Public Health and an executive advisor in healthcare to the technology innovation hub AI and Robotics Tech Park at IISc. He co-chairs the technology workstream at the Lancet Citizens Commission on reimagining India’s health systems. In his work with nonprofits on orphan disease outreach, Vijay is driven by the mission to achieve “no disease orphan” by 2030.
As an entrepreneur, Vijay led Strand Life Sciences, India’s leading Genomics and Precision Medicine Solutions company, as Executive Chairman (2000-2018). He has co-founded several other companies in handheld hardware (Picopeta Simputers), financial technology (Yantri Labs – algorithmic trading), and biotechnology (CRISPRBITS) and serves as a mentor to the deep science and tech innovation ecosystem in Bangalore.
We’ll discuss a systematic approach to Prioritized Selection Problems, in which an organization is presented with a set of individuals, and must choose a subset.
We assume that the organization has a complete priority ranking of the individuals. In addition, the selection rule may be constrained to select from a predefined collection of feasible subsets, which may be specified explicitly or implicitly. We identify a simple family of rules and characterize it using some natural axiomatic properties. In cases where the feasible subsets are implicitly specified through upper and lower quotas or through reserves, we identify special cases for which there exists a selection that unambiguously selects higher-priority individuals than any other feasible selection.
Furthermore, we identify conditions under which this selection can be found in polynomial time, and conditions in which doing so is computationally hard.
This class of problems is useful in assigning and matching models with distributional objectives/constraints. Our approach is motivated by the observation that it is easier for the general public and policymakers to reason about outcomes rather than algorithms.
[Joint work with Nick Arnosti and Carlos Bonet]
Jay Sethuraman is a Professor of Industrial Engineering and Operations Research at Columbia University, is currently the chair of the IEOR department, and is a proud graduate of the CSA Department at the Indian Institute of Science.
His research interests are in market design, discrete optimization and its applications, scheduling theory, and applied probability.
Balaraman Ravindran, IIT Madras
Bharath Kumar, Sensara Technologies
Manish Gupta, Google Research
Neeldhara Mishra, IIT Gandhinagar
Ram D. Sriram, National Institute of Standards and Technology, USA
Siddhartha Gadgil, IISc
Moderator: Praphul Chandra, Atria University
In this talk, I will describe a close connection between discrete dynamical systems and games and use this close connection to discuss convergence properties of dynamics in coordination and anti-coordination games. Some connections with evolutionary game dynamics will also be discussed.
Madhav Marathe is a Distinguished Professor in Biocomplexity, the Executive Director of the Biocomplexity Institute, and a Professor in the Department of Computer Science at the University of Virginia (UVA). His research interests are in network science, Sustainable habitats, AI, foundations of computing. Over the last 25 years, his division has supported federal and state authorities in their effort to respond to a number of problems arising in the context of national security, sustainability and pandemic science, including, the COVID-19 pandemic.
Before joining UVA, he held positions at Virginia Tech and the Los Alamos National Laboratory. He is a Fellow of the IEEE, ACM, SIAM and AAAS
The widespread use of sensors and internet of things (IoT) has enabled machine learning (ML) models to learn and analyze vast amount of data, make predictive inferences leading to actionable decisions, and fine tune system parameters for performance optimization. ML over IoT supports a variety of society-scale applications, such as smart homes/cities, smart grid, smart health, smart agriculture, smart transportation, smart manufacturing, and so on. However, the integration of ML over resource-constraint IoT devices poses significant research challenges owing to the uncertainty, scalability, heterogeneity, data quality, security, privacy, and fairness issues.
This talk will explore novel frameworks and models to address three specific problems: (i) handling uncertainty (noisy labels) in federated learning (FL) over IoT
(ii) securing FL models via malicious client detection
(iii) achieving performance fairness in FL under attacks
The underlying principles will be based on information-theoretic divergence, novel gradient space, and regularizing techniques. Real-world case studies and experimental performance will be presented, where possible, with directions of future research.
Sajal K. Das is a Curators’ Distinguished Professor and Daniel St. Clair Endowed Chair in Computer Science at Missouri University of Science and Technology, Rolla, where he was the Chair of Computer Science Department during 2013-2017. He also served the National Science Foundation (NSF) as a Program Director in the Computer and Network Systems Division. His interdisciplinary research spans cyber-physical systems, IoT, cybersecurity, machine learning and data science, wireless and sensor networks, mobile and pervasive computing, smart environments, parallel/cloud/edge computing, social and biological networks, applied graph theory and game theory.
He has contributed significantly to these areas and published extensively in top-tier venues (more than 350 journal articles and more than 450 peer-reviewed conference papers). He coauthored four books, 59 book chapters, and 5 US patents and directed over $24 million funded research projects. His h-index is 99 with more than 42,000 citations. He is the founding Editor-in-Chief of Elsevier’s Pervasive and Mobile Computing journal (since 2005) and an Associate Editor of IEEE Transactions on Dependable and Secure Computing, IEEE Transactions on Mobile Computing, IEEE Transactions on Sustainable Computing, IEEE/ACM transactions on Networking, and ACM Transactions on Sensor Networks. A founder of IEEE PerCom, WoWMoM, SMARTCOMP and ACM ICDCN conferences, he has served as General and Program Chair of numerous conferences and workshops. He is a recipient of 12 Best Paper Awards in flagship conferences like ACM MobiCom and IEEE PerCom, and awards of excellence for teaching, mentoring and research including the IEEE Computer Society’s Technical Achievement Award and University of Missouri System President’s Award for Sustained Career Excellence.
He has graduated 12 postdoctoral fellows, 51 Ph.D. scholars, 31 MS thesis, and numerous undergraduate research students. He is an Infosys Visiting Chair Proessor at IISc, a Distinguished alumnus of the Indian Institute of Science, Bangalore and a Fellow of the IEEE, National Academy of Inventors (NAI) and Asia-Pacific Artificial Intelligence Association (AAIA).
It was recently shown that there are games in which no dynamics converge to the Nash equilibria. I will discuss two research streams motivated by this negative result: An investigation of complexity results seeking to sharpen this theorem, and a new conception of the meaning of the game that is both well motivated and tractable.
Christos Papadimitriou is the Donovan Family Professor of Computer Science at Columbia University. Before Columbia, he taught at UC Berkeley for 22 years and before that at Harvard, MIT, National Technical University of Athens, Stanford and UCSD. His work is about understanding mathematically the nature and limitations of computation and using computation as a lens for making progress in scientific problems in fields such as control theory, economics and game theory, evolution and neuroscience.
He holds a Ph.D. from Princeton University and honorary doctorates from nine universities, including ETH, EPFL and the universities of Athens, Cyprus and Paris Dauphine. He is a member of the National Academy of Sciences, the National Academy of Engineering and the American Academy of Arts and Sciences of the US, and he has received the Donald E. Knuth Prize, the Gödel Prize, the IEEE John von Neumann Medal, the IEEE Computer Society Women of ENIAC Computer Pioneer Award, the John von Neumann Theory Prize, the IEEE SC Charles Babbage Award and the Harvey Prize from Technion. He has also written three novels, including a New York Times bestseller.
Rakesh Agrawal is the President and Founder of the Data Insights Laboratories, San Jose, USA. He is a member of the National Academy of Engineering, both USA and India, a Fellow of ACM, and a Fellow of IEEE. He has been both an IBM Fellow and a Microsoft Fellow. He has served as visiting professor internationally (EPFL-Switzerland, Kyoto University-Japan, Indian Institute of Science-Bangalore, IIT-Bombay).
Rakesh has published 200+ highly influential papers, including the 1st and 2nd highest cited in databases and data mining. They have been cited 140,000+ times with 35+ of them receiving 500+ citations and 3 receiving 8000+ citations (Google Scholar). He has been issued 88 patents. For his ground-breaking research, he has received the Innovation Awards given to the topmost researchers from two ACM-SIGs: SIGKDD and SIGMOD. In addition, his papers have received six test-of-time awards from five conferences: SIGMOD (twice), VLDB, ICDE, EDBT, WSDM. These awards recognize papers published ten years back that had the most influence in the field and industry.
Rakesh’s research has a far-reaching impact on commercial products and services. For instance, IBM’s Intelligent Miner grew straight out of Rakesh’s data mining research, which also influenced the products of several companies (e.g., Oracle, SAP, SAS, SPSS, WEKA). He pioneered key concepts in data privacy, including Hippocratic database, privacy-preserving data mining, and sovereign information sharing, which have influenced data governance and compliance. He devised techniques for mining workflows from the logs of activities which have been used in commercial products like Flowmark. He invented techniques for automatically organizing and presenting unstructured information and architected their use for building catalogs for Bing’s Ciao product search. He formulated the foundational principles for diversified ranking of search results which shaped the SIGIR TREC’s diversity task.
Rakesh has played key roles in projects of significant societal benefits (e.g., 2005 study on Improving Education System for the President of India, 2006 NRC study of Voter Registration Databases, 2009 NRC study of S&T strategies of six countries). Rakesh applied his technology for enriching textbooks to the NCERT books used by millions in India and has provided the results to NCERT to enable the authors to incorporate the improvements in future editions.
Is it possible to construct mechanisms that ensure the dominance of dynamic truth-telling for agents comprised of stochastic dynamic systems? We show that it is possible to achieve subgame perfect dominance of truth-telling for the special case of LQG agents if system parameters are known and agents are rational. This is accomplished through the construction of a sequence of layered payments over time that decouples the intertemporal effect of current bids on future net utilities. An important motivating example arises in power systems where a System Operator has to ensure the balance of generation and consumption at all times. We show that there is a modified “Scaled” VCG (SVCG) mechanism that satisfies incentive compatibility, social efficiency, budget balance, and individual rationality, under a certain “market power balance” condition where no agent is neither too negligible nor too dominant. The SVCG payments converge to the Lagrange payments that correspond to the true price in the absence of strategic considerations, as the number of agents in the market increases. [Joint work with Ke Ma].
P. R. Kumar obtained a B. Tech degree from IIT Madras (1973), and D.Sc. from Washington Univ, St. Louis (1977). He served in the Math Dept at Univ. of Maryland Baltimore County (1977-84), and ECE and CSL at Univ. of Illinois, Urbana-Champaign (1985-2011), before joining Texas A&M Univ, where he is a University Distinguished Professor, a Regents Professor, and holds the O’Donnell Chair-I.
He is a member of the U.S. NAE, The World Academy of Sciences, and the Indian NAE. He was awarded a Doctor Honoris Causa by ETH, Zurich. He received the IEEE Alexander Graham Bell Medal, IEEE Field Award for Control Systems, Eckman Award of AACC, Ellersick Prize of IEEE ComSoc, Outstanding Contribution Award of ACM SIGMOBILE, Infocom Achievement Award, and SIGMOBILE Test-of-Time Paper Award.
He is a Fellow of IEEE, ACM, and IFAC. He is an Honorary Professor at IIT Hyderabad. He was awarded a Distinguished Alumnus Award from IIT Madras, an Alumni Achievement Award from Washington Univ, and a Drucker Eminent Faculty Award from UIUC. His current focus includes CPS, Security, UTM, Wireless Networks, ML, and Power Systems.
Proactive maintenance in manufacturing industries relies heavily on the accurate diagnosis and prognosis of cutting tools. This paper presents a robust prognostics and health management (PHM) framework tailored for cutting operations on computer numerical control (CNC) machines. Our framework, designed to support proactive decision-making in tool maintenance, estimates tool wear and remaining useful life (RUL) using sensor-derived indicators of tool health. The integration of machine learning techniques and physics-based methods enhances the precision of our tool wear and RUL predictions, making our framework adaptable to various machining conditions. Validation through a series of run-to-failure tests across diverse cutting settings and machines underscores the versatility and robustness of our approach. The results reveal significant improvements in predictive accuracy and operational efficiency, demonstrating the framework’s potential to optimize maintenance schedules, reduce downtime, and improve industrial processes.
Krishna R. Pattipati is the Distinguished Professor Emeritus and the Collins Aerospace Chair Professor of Systems Engineering in the Department of Electrical and Computer Engineering at the University of Connecticut, Storrs, CT, USA. Prof. Pattipati’s research activities are in the areas of proactive decision support, autonomy, and optimization-based learning and inference. A common theme among these applications is that they are characterized by a great deal of uncertainty, complexity, and computational intractability. He has published over 580 scholarly journals and conference papers in these areas. He is a co-founder of Qualtech Systems, Inc., a firm specializing in advanced integrated diagnostics software tools (TEAMS, TEAMS-RT, TEAMS-RDS, TEAMATE, PackNGo), and serves on the board of Aptima, Inc.
Prof. Pattipati received the Centennial Key to the Future award in 1984 from the IEEE Systems, Man, and Cybernetics (SMC) Society, and was elected a Fellow of the IEEE in 1995 for his contributions to discrete-optimization algorithms for large-scale systems and team decision-making. He received the Andrew P. Sage Award for the Best SMC Transactions Paper for 1999, the Barry Carlton Award for the Best Aerospace and Electronic Systems (AES) Transactions Paper for 2000, the 2002 and 2008 NASA Space Act Awards for “A Comprehensive Toolset for Model-based Health Monitoring and Diagnosis,” and “Real-time Update of Fault-Test Dependencies of Dynamic Systems: A Comprehensive Toolset for Model-Based Health Monitoring and Diagnostics”, the 2003 AAUP Research Excellence Award, the 2005 School of Engineering Teaching Excellence Award at the University of Connecticut, and the 2023 Distinguished Alumnus Award from the Indian Institute of Technology, Kharagpur. Prof. Pattipati served as Editor-in-Chief of the IEEE Transactions on SMC-Cybernetics (Part B) from 1998 to 2001.
Prof. Krishna R. Pattipati
Department of Electrical and Computer Engineering
University of Connecticut (UCONN)
Storrs, CT 06269-2157
krishna.pattipati@uconn.edu
In OFDMA cellular systems (e.g., 5G), the scheduler is the mechanism that provides quality to the traffic flows over the high-speed, finely slotted OFDMA transport. For bulk elastic transfers (such as TCP file transfers), resource sharing is formulated as a utility optimization problem over the throughput region. When user utility is the log of the user throughput, we get proportional fair (PF) sharing. Restricting attention to downlink flows, with the full-buffer assumption, we have studied an extension of PF scheduling that accounts for the probability of losing the data block that is transmitted. This gives rise to the PF(LA) scheduler (proportional fair, loss adaptive), where the PF scheduling indices are modified by the expected number of reattempts of the scheduled data block.
For providing rate guarantees, we develop a two-time scale scheduler (PF-RG-LM: proportional fair, rate guarantee, Lagrange multiplier), which performs accurately as compared to a recently reported single time scale algorithm. The PF(LA)-RG-LM algorithm combines the above techniques to provide a rate guarantee while being loss-adaptive. We will discuss the formulation and the algorithms, provide partial theory, and show over-the-air experimental results from the OAI-based 5G private network in the 5G Lab in the ECE Department, IISc.
Prof. Anurag Kumar (B.Tech (1977) IIT Kanpur, PhD (1981) Cornell Univ.) was a Member of Technical Staff in AT&T Bell Laboratories (1981-1988), before returning to India and joining the Indian Institute of Science (IISc) as a faculty member in the ECE Department.
He became a Professor in 1996, the Director of IISc during 2014-2020, and an Honorary Professor after his superannuation in 2020. Since 1st January 2024, he is an Indian National Science Academy (INSA) Distinguished Professor.
(joint with Mohsen Pourponeh (University of Maastricht), Rasoul Ramezanian (University of Lausanne), and Vilok Taori (ISI Delhi).
We consider a variant of the Assignment Game model of Shapley and Shubik (1971) where the matching of workers and firms is commonly observed but the division of surplus between a matched pair is privately observed only by the firm and the worker concerned. We define and apply a stability notion in this setting based on the iterative elimination of blocked matching outcomes which captures the idea that the absence of blocking pairs conveys no further information regarding the payoffs received by the other participants in the market. We show that the set of private payment stable matching outcomes always exists, the corresponding assignments are efficient in terms of maximizing total surplus, and the set is characterized by the worker optimal and firm optimal stable matching outcomes.
Arunava Sen is a Professor Emeritus at the Indian Statistical Institute, Delhi Centre.
He obtained BA and MA degrees from Delhi University, an MPhil from Oxford University, and a Ph.d from Princeton University. His research interests lie in game theory, mechanism design, and market design theory. He is a Fellow of the Econometric Society, an Economic Theory Fellow, and a Fellow of the Game Theory Society.
He is a recipient of the Infosys Prize in Social Sciences, the Mahalonobis Medal, and the AL Nagar Fellowship of the Indian Econometric Society.
After recalling how Sir Peter Whittle motivated his eponymous index from Lagrange multipliers, I shall introduce the `Lagrange index’ (equivalent to the `LP index’) which is usually simpler to compute and can outperform the Whittle index. (Joint work with Konstantin Avrachenkov and Pratik Shah)
Prof. Vivek Borkar obtained his B.Tech. in Elec. Engg. from IIT Bombay, M.S. in Systems and Control Engg. from Case Western Reserve Uni., and Ph.D. from the Uni. of California at Berkeley in Elec. Engg. and Computer Science, in ’76, ’77, and ’80, resp. After a visiting stint at the Uni. of Twente (1980-81) in the Netherlands, he
held positions at the TIFR CEntre for Applicable Mathematics and Indian Institute of Science in Bengaluru, and TIFR and IIT Bombay in Mumbai. He is currently an Emeritus Fellow at IIT Bombay. He is a Fellow of IEEE, AMS, TWAS, and various science and engineering academies in India.
His research interests are stochastic optimization and control, covering theory, algorithms, and applications, particularly to communications.
Prof. V. S. Borkar
Department of Electrical Engineering,
IIT, Powai, Mumbai 400076, India.
Email: borkar.vs(at)gmail.com
With advances in technology, it has become possible to extract valuable biological information at the granularity of single cells. Single cell data for up to a million observations and with thousands of measurements per cell are now available. Clustering such data can help uncover a host of new information on cell types, cellular signatures and multiple cellular processes. However, given that single cells data are large, high-dimensional, noisy and possibly contain small but informative groups of cells, the algorithms have to be designed with care.
In this talk, we will first provide quick introductions to basic molecular biology and single cell RNA-seq data. This will be followed by a description of an algorithm for clustering single cell RNA-seq data. Finally, a few experimental results and their biological insights will be discussed.
We consider the problem of allocating items/goods to agents when the agents have submodular valuation functions. Such problems arise in a number of areas including algorithmic game theory. Continuous relaxations of submodular functions have played a key role in providing improved algorithms for several problems including the welfare maximization problem.
We will briefly review this result and describe a new simple lemma that has applications in the max-min setting. Time permitting we will also describe how continuous relaxations also help in understanding the max-min share problem that arises in fair allocation.
Chandra Chekuri is the Paul and Cynthia Saylor Professor in the Department of Computer Science at the University of Illinois, Urbana-Champaign. He joined the university in 2006 after spending eight years at Lucent Bell Labs. Prior to that, he received his PhD from Stanford University in 1998, and an undergraduate degree from IIT Chennai in 1993.
He is interested in the design and analysis of algorithms, combinatorial optimization, and theoretical computer science. He recently became an ACM Fellow for his contributions to approximation algorithms and submodular optimization.
Prof. Manoj Kumar Tiwari (FNAE, FASc, FNASc, and FIISE) has been the Founding Director of IIM Mumbai since August 2023, where he also served as Director of the National Institute of Industrial Engineering (NITIE) Mumbai from 2019-2023. He was a Professor with a Higher Academic Grade (HAG) in the Department of Industrial and Systems Engineering at the Indian Institute of Technology, Kharagpur. He is actively involved in research relevant to the application of optimization, modeling, decision support systems, and data mining in logistics, supply chain management, and manufacturing research domains. He is associate/senior editor of several renowned journals like IJPR, POMS, JIM, IEEE-SMCA, INS, and CAIE.
Prof Tiwari has led numerous projects and consultancy assignments with prominent industry and government organizations in India, including Indian Air Force, Procter and Gamble, TATA Hitachi, and ports located in eastern India. In addition, he aggressively engaged in collaborative research with partners from Loughborough University UK, Institute for Manufacturing (IfM) at the University of Cambridge UK, Warwick Manufacturing Group UK, Durham University UK, and the University of Connecticut USA.
Prof. Tiwari is an author of more than 360 articles in leading international journals with an H index-86. He is a fellow of The National Academy of Sciences India (NASI), Indian Academy of Sciences (FASc), and Indian National Academy of Engineering (INAE), the Institute of Industrial and Systems Engineers (IISE), USA, and also recognized with David F. Baker Distinguished Research Award (IISE, USA).