arXiv research

We introduce CSE for MLSF games and devise online learning algorithms for achieving no-external Stackelberg-regret.

The online problem of computing the top eigenvector is fundamental to machine learning. In both adversarial and stochastic settings, previous results (such as matrix multiplicative weight update, follow the regularized leader, follow the compressed leader, block power method) either achieve optimal regret but run slow,…

CB-RL solves complex decision-making problems with contextual information and exogenous events.

Characterizes preferences for decision-making under uncertainty using a leader-follower game model.

Investors with asymmetric information play a game to optimize their portfolios.

New RL algorithms learn QSE from strategic feedbacks with sample efficiency.

We consider distributed optimization under communication constraints for training deep learning models. We propose a new algorithm, whose parameter updates rely on two forces: a regular gradient step, and a corrective direction dictated by the currently best-performing worker (leader). Our method differs from the param…

Bitcoin volatility shows multifractal structure, contradicting rough volatility models.

Distillation is an effective knowledge-transfer technique that uses predicted distributions of a powerful teacher model as soft targets to train a less-parameterized student model. A pre-trained high capacity teacher, however, is not always available. Recently proposed online variants use the aggregated intermediate pr…

Using a two-point correlation technique, we study emergence of market efficiency in the emergent Russian futures market by focusing on lagged correlations. The correlation strength of leader-follower effects in the lagged inter-market correlations on the hourly time frame is seen to be significant initially (2009-2011)…

Study of $2 imes 2$ zero-sum games with noisy observations and commitments.

Transformers cluster meaningless words around leaders for sentiment analysis.

FPML algorithm reduces regret by limiting the number of arms chosen per round.

Neural operators approximate Stackelberg game solutions.

Game theory approach to predicting and responding to interventions based on causal relationships.

Deep learning models predict financial market trends from social media leaders.

Mutually interacting components form complex systems and the outputs of these components are usually long-range cross-correlated. Using wavelet leaders, we propose a method of characterizing the joint multifractal nature of these long-range cross correlations, a method we call joint multifractal analysis based on wavel…

New algorithm for online optimization over symmetric cones, unifying previous methods.

SLHF uses sequential game theory to optimize preferences from human feedback.

FTPL with Fréchet perturbation achieves near optimal regret bounds for m-set semi-bandit problems.

We study the problem of online learning with non-convex losses, where the learner has access to an offline optimization oracle. We show that the classical Follow the Perturbed Leader (FTPL) algorithm achieves optimal regret rate of $O(T^{-1/2})$ in this setting. This improves upon the previous best-known regret rate of…

Adaptive learning rates improve FTPL's BOBW guarantees in bandit problems.

New mechanism designs regulate herding in financial markets.

Improved theoretical guarantees for Top Two algorithms.

Diestel-Leader graphs are neither hyperbolic nor CAT(0), so their visual boundaries may be pathological. Indeed, we show that for $d>2$, $\partial\text{DL}_d(q)$ carries the indiscrete topology. On the other hand, $\partial\text{DL}_2(q)$, while not Hausdorff, is $T_1$, totally disconnected, and compact. Since $\text{D…

We revisit the fundamental problem of prediction with expert advice, in a setting where the environment is benign and generates losses stochastically, but the feedback observed by the learner is subject to a moderate adversarial corruption. We prove that a variant of the classical Multiplicative Weights algorithm with …

Paper optimizes FTPL for adversarial and stochastic bandits with specific tail distributions.

Model improves BERT for answering multiple-choice questions in large texts.

Paper tackles continual reinforcement learning by forgetting, proposing a planning method with online world models.

In this paper, we provide a novel construction of the linear-sized spectral sparsifiers of Batson, Spielman and Srivastava [BSS14]. While previous constructions required $Ω(n^4)$ running time [BSS14, Zou12], our sparsification routine can be implemented in almost-quadratic running time $O(n^{2+\varepsilon})$. The funda…

Paper finds political networks reduce bond issuance costs in China.

We analyze linear McKean-Vlasov forward-backward SDEs arising in leader-follower games with mean-field type control and terminal state constraints on the state process. We establish an existence and uniqueness of solutions result for such systems in time-weighted spaces as well as a {convergence} result of the solution…

Game-theoretic models of learning are a powerful set of models that optimize multi-objective architectures. Among these models are zero-sum architectures that have inspired adversarial learning frameworks. An important shortcoming of these zeros-sum architectures is that gradient-based training leads to weak convergenc…

Advances FTPL results for bandit problems with unbounded perturbations.

FTPL policy achieves best-of-both-worlds regret in decoupled bandits with reduced computational cost.

Given a bordified space, Karlsson defines an incidence geometry of stars at infinity. These stars and their incidence are closely related to well-understood objects when the space is hyperbolic, CAT(0), or a bounded convex domain with the Hilbert metric. A question stemming from Karlsson's original paper was whether or…

Paper analyzes FTPL's effectiveness in combinatorial semi-bandit problems.

Paper studies zero-sum games with noisy observations and identifies equilibrium conditions.

New algorithm optimizes multi-armed bandits with low computational cost.

We fill a void in merging empirical and phenomenological characterisation of the dynamical phase transitions in complex systems by identifying three of them on real-life financial markets. We extract and interpret the empirical, numerical, and semi-analytical evidences for the existence of these phase transitions, by c…

Symbolic Data Analysis is based on special descriptions of data - symbolic objects (SO). Such descriptions preserve more detailed information about units and their clusters than the usual representations with mean values. A special kind of symbolic object is a representation with frequency or probability distributions …

Paper optimizes multi-agent learning in Markov games with generative model.

Black-Scholes (BS) is the standard mathematical model for option pricing in financial markets. Option prices are calculated using an analytical formula whose main inputs are strike (at which price to exercise) and volatility. The BS framework assumes that volatility remains constant across all strikes, however, in prac…

Network analysis reveals changing cryptocurrency market leaders.

This paper improves FTPL algorithm for semi-bandit problems with best-of-both-worlds guarantees.

New aggregation strategy handles unbounded losses with regret bounds.

Estimates network structure and interaction rules from multiple agent trajectories.

New algorithms improve dueling bandit performance in multiplayer settings.