Abstract and link to the full paper for my publications and preprints.
Subhamon Supantha, A. Sinha
Work done while at IIT Bombay and Tata Institute of Fundamental Research, Mumbai (TIFR)
Abstract: We consider a generalization of the celebrated Online Convex Optimization (OCO) framework with online adversarial constraints. We present two algorithms having simple modular structures that yield universal dynamic regret and cumulative constraint violation bounds, improving upon the state-of-the-art results. Our results hold in the most general case when both the cost and constraint functions are chosen arbitrarily by an adversary, and the constraint functions need not contain any common feasible point. The results are established by reducing the constrained learning problem to an instance of the standard OCO problem with specially constructed surrogate cost functions.
Accepted at the 37th International Conference on Algorithmic Learning Theory, 2026 (ALT 2026).
Received ALT 2026 Elegance Award!
D. Sarkar, A. Chakrabartty, Subhamon Supantha, P. Dey, A. Sinha
Work done at TIFR, Mumbai
Abstract: We study a generalization of the Online Con- vex Optimization (OCO) framework with time- varying adversarial constraints. In this problem, after selecting a feasible action from the con- vex decision set X, a convex constraint function is revealed alongside the cost function in each round. Our goal is to design a computationally efficient learning policy that achieves a small re- gret with respect to the cost functions and a small cumulative constraint violation (CCV) with re- spect to the constraint functions over a horizon of length T. It is well-known that the projection step constitutes the major computational bottle- neck of the standard OCO algorithms. However, for many structured decision sets, linear func- tions can be efficiently optimized over the deci- sion set. We propose a projection-free online pol- icy which makes a single call to a Linear Program (LP) solver per round. Our method outperforms state-of-the-art projection-free online algorithms with adversarial constraints, achieving improved bounds of ˜O(T³⁄⁴) for both regret and CCV. The proposed algorithm is conceptually simple - it first constructs a surrogate cost function as a non- negative linear combination of the cost and con- straint functions. Then, it passes the surrogate costs to a new, adaptive version of the online con- ditional gradient subroutine, which we propose in this paper.
Accepted at the 29th Annual Conference on Artificial Intelligence and Statistics (AISTATS 2026).
Subhamon Supantha, N. K. Sharma
Work done at University of Hyderabad
Abstract: We have used agent-based modeling as our numerical method to artificially simu- late a dynamic real economy where agents are rational maximizers of an objective function of Cobb-Douglas type. The economy is characterised by heterogeneous agents, acting out of local or imperfect information, monopolistic competition, perfect product differentiation, allowance for increasing returns to scale tech- nology and trade in disequilibrium. An algorithm for economic activity in each period is devised and a general purpose open source agent-based model is devel- oped which allows for counterfactual inquiries, testing out treatments, analysing causality of various economic processes, outcomes and studying emergent prop- erties. 10,000 simulations, with 10 firms and 80 consumers are run with varying parameters and the results show that from only a few initial conditions the econ- omy reaches equilibrium while in most of the other cases it remains in perpetual disequilibrium. It also shows that from a few initial conditions the economy reaches a disaster where all the consumer wealth falls to zero or only a single pro- ducer remains. Furthermore, from some initial conditions, an ideal economy with high wage rate, high consumer utility and no unemployment is also reached. It was also observed that starting from an equal endowment of wealth in consumers and in producers, inequality emerged in the economy. In majority of the cases most of the firms(6-7) shut down because they were not profitable enough and only a few firms remained. Our results highlight that all these varying outcomes are possible for a decentralized market economy with rational optimizing agents.
Accepted (presentation), International Conference on Computing in Economics and Finance (2025).
Subhamon Supantha, P. Pandit, A. Rangamani
Work done while at IIT Bombay
Abstract: We revisit the phenomenon of neural collapse under the case where the embedding dimension cannot support a simplex ETF for all the classes, but the classes have a hierarchy which aid a different kind of geometry in which the embeddings concentrate. Our result shows that the average of embeddings from the same super class (called super-class means) form a simplex ETF, furthermore, the sub-class mean embeddings from the same super-class, form another simplex ETF in the plane orthogonal to the origin super-class mean.