A locally-built, LLM-digested index of recent arXiv papers in quant finance, geometry/topology, and statistical ML — keyword search served straight from SQLite on this machine.
Path-independent equilibrium models improve network performance on harder problems.
Hard instances, which require a long time for a specific algorithm to solve, help (1) analyze the algorithm for accelerating it and (2) build a good benchmark for evaluating the performance of algorithms. There exist several efforts for automatic generation of hard instances. For example, evolutionary algorithms have b…
This paper poses some basic questions about instances (hard to find) of a special problem in 3-manifold topology. "Important though the general concepts and propositions may be with the modern industrious passion for axiomatizing and generalizing has presented us...nevertheless I am convinced that the special problems …
This paper introduces self-paced task selection to multitask learning, where instances from more closely related tasks are selected in a progression of easier-to-harder tasks, to emulate an effective human education strategy, but applied to multitask machine learning. We develop the mathematical foundation for the appr…
Framework for few-shot relation classification with minimal training data.
We propose a general framework for sequential and dynamic acquisition of useful information in order to solve a particular task. While our goal could in principle be tackled by general reinforcement learning, our particular setting is constrained enough to allow more efficient algorithms. In this paper, we work under t…
We study the problem of interactively learning a binary classifier using noisy labeling and pairwise comparison oracles, where the comparison oracle answers which one in the given two instances is more likely to be positive. Learning from such oracles has multiple applications where obtaining direct labels is harder bu…
The analysis of continously larger datasets is a task of major importance in a wide variety of scientific fields. In this sense, cluster analysis algorithms are a key element of exploratory data analysis, due to their easiness in the implementation and relatively low computational cost. Among these algorithms, the K -m…
Machine learning predicts flight connections for airline crew scheduling.
Using methods of descriptive theory it is shown that the classification problem for wild knots is strictly harder than that for countable structures.
The easiness at which adversarial instances can be generated in deep neural networks raises some fundamental questions on their functioning and concerns on their use in critical systems. In this paper, we draw a connection between over-generalization and adversaries: a possible cause of adversaries lies in models desig…
We give a generalisation of the theory of optimal destabilizing 1-parameter subgroups to non-algebraic complex geometry. Consider a holomorphic action of a complex reductive Lie group on a finite dimensional (possibly non-compact) Kähler manifold . Using a Hilbert type criterion for the (semi)st…
SATNet solves the Symbol Grounding Problem, enabling self-supervised learning.
We show that the ``classical'' Harder-Narasimhan filtration associated to a non semistable vector bundle can be viewed as a limit object for the action of the gauge group in the direction of an optimal destabilizing vector. This vector appears as an extremal value of the so called "maximal weight function". We give…
This paper explores the complexity of learning representations in contextual linear bandits.
Study slopes of direct images in complex manifolds, proving a Mehta-Ramanathan type theorem.
Professional software developers spend a significant amount of time fixing builds, but this has received little attention as a problem in automatic program repair. We present a new deep learning architecture, called Graph2Diff, for automatically localizing and fixing build errors. We represent source code, build config…
New algorithms tackle robust RL with linear models, revealing unique challenges.
Paper shows FB and FC are equally hard up to logarithmic factors.
Optimizes reinforcement learning by prioritizing sets of samples over individual ones.
This paper poses some basic questions about instances (hard to find) of a special problem in 3-manifold topology. "Important though the general concepts and propositions may be with the modern industrious passion for axiomatizing and generalizing has presented us ... nevertheless I am convinced that the special problem…
New method shows multi-objective bandits are not harder than single-objective ones.
We address the problem of semi-supervised learning in relational networks, networks in which nodes are entities and links are the relationships or interactions between them. Typically this problem is confounded with the problem of graph-based semi-supervised learning (GSSL), because both problems represent the data as …
Generating datasets that "look like" given real ones is an interesting tasks for healthcare applications of ML and many other fields of science and engineering. In this paper we propose a new method of general application to binary datasets based on a method for learning the parameters of a latent variable moment that …
Study shows sample complexity for learning optimal policies in SSP with generative model.
We show that closed arithmetic hyperbolic n-dimensional orbifolds with larger and larger volumes give rise to triangulations of the underlying spaces whose 1-skeletons are harder and harder to embed nicely in Euclidean space. To show this we generalize an inequality of Gromov and Guth to hyperbolic n-orbifolds and find…
The paper tackles fair sequential decision making with biased linear bandit feedback.
Adaptive compute allocation improves model performance by prioritizing harder queries.
This is the first of a series of papers where we relate tangent cones of Hermitian-Yang-Mills connections at an isolated singularity to the complex algebraic geometry of the underlying reflexive sheaf, when the sheaf is locally modelled on the pull-back of a holomorphic vector bundle from the projective space. In this …
This survey intends to present the basic notions of Geometric Invariant Theory (GIT) through its paradigmatic application in the construction of the moduli space of holomorphic vector bundles. Special attention is paid to the notion of stability from different points of view and to the concept of maximal unstability, r…
New RLHF algorithm identifies optimal policies from human feedback without explicit reward inference.
Inspired by Katz-Mazur theorem on crystalline cohomology and by Eskin-Kontsevich-Zorich's numerical experiments, we conjecture that the polygon of Lyapunov spectrum lies above (or on) the Harder-Narasimhan polygon of the Hodge bundle over any Teichmüller curve. We also discuss the connections between the two polygons a…
The goal of this paper was to predict the placement in the multiplayer game PUBG (playerunknown battleground). In the game, up to one hundred players parachutes onto an island and scavenge for weapons and equipment to kill others, while avoiding getting killed themselves. The available safe area of the game map decreas…
New method predicts language model scaling risks with limited data.
In stochastic multi-armed bandits, the reward distribution of each arm is assumed to be stationary. This assumption is often violated in practice (e.g., in recommendation systems), where the reward of an arm may change whenever is selected, i.e., rested bandit setting. In this paper, we consider the non-parametric rott…
The paper tackles adaptive policy selection to maximize social welfare, achieving optimal regret bounds.
Paper studies convergence of Yang-Mills-Higgs flow on Kähler manifolds.
Model learns brevity by exposing to easy problems, improving efficiency without explicit length penalties.
The paper studies obstructions to solutions of the Wess-Zumino-Witten equation and its generalizations.
New algorithm reduces regret in asynchronous multiplayer bandits to constant or logarithmic levels.
We suggest a new algorithm for finding a canonical representative of a given braid, and also for the harder problem of finding a -consistent representative. We conjecture that the algorithm is quadratic-time. We present numerical evidence for this conjecture, and prove two results: (1) The algorithm terminates in …
Study describes splitting and filtration of Hodge bundle on quadratic differentials.
The research proposes a stopping rule for reinforcement learning algorithms based on instance-dependent confidence.
In this work we establish the first linear convergence result for the stochastic heavy ball method. The method performs SGD steps with a fixed stepsize, amended by a heavy ball momentum term. In the analysis, we focus on minimizing the expected loss and not on finite-sum minimization, which is typically a much harder p…
Reasoning models generate differently based on problem difficulty, not just length.
Expressive quantum circuits are harder to train due to flatter cost landscapes.
Recently neural networks and multiple instance learning are both attractive topics in Artificial Intelligence related research fields. Deep neural networks have achieved great success in supervised learning problems, and multiple instance learning as a typical weakly-supervised learning method is effective for many app…
RLVR training dynamics reveal an implicit curriculum that shapes learning progression.