The paper presents Tahoe, a system that optimizes Text-to-SQL performance through dynamic data management and hint learning.

This paper develops a policy-learning framework to optimally assign prediction tasks to multiple agents, considering individual agent expertise and capacity constraints, achieving systematic performan…

This paper introduces survey sampling techniques to estimate or minimize empirical pairwise loss functions, showing that targeting informative pairs significantly reduces computational cost while main…

The paper introduces eXTC, a novel framework that combines structured prompt optimization, knowledge distillation, and reinforcement learning to create a highly performant and fully interpretable text…

The paper analyzes the sample complexity of contextual bandits in the $s$-sparse setting, achieving optimal sample bounds for identifying an $\epsilon$-optimal policy.

Prompt Codebooks (PCO) introduces a compositional framework that treats prompt optimization as discrete learning over reusable instruction units, significantly improving LLM performance while drastica…

The paper proposes a novel temporal and structural credit assignment framework to efficiently optimize multi-agent LLM systems by decomposing the error signal and using targeted, discrete gradient upd…

The paper theoretically analyzes the properties that optimal sparse autoencoder (SAE) dictionaries must satisfy, deriving constraints that explain observed SAE behaviors like hierarchical splitting an…

The paper theoretically explains that optimizing LLMs solely on outcomes leads to brittle reasoning (Reward-Induced Manifold Collapse) by favoring low-complexity shortcuts, and proposes process-based…

The paper formalizes the problem of representation identifiability in supervised learning, showing that a representation property is identifiable if and only if it is constant across all possible fact…

The paper introduces trust functions to filter weak supervision labels, enabling near-lossless weak-to-strong generalization by selectively training a strong student using only the most reliable weak…

The paper introduces PRAXIS, a novel algorithm that efficiently approximates the computation of 'Rashomon sets' for decision trees, significantly reducing memory and runtime complexity.

The paper introduces Feedback Distillation, a novel training method that uses a language model's privileged feedback to provide token-level supervision, significantly improving complex reasoning tasks…

The paper establishes tight upper and lower bounds on the statistical cost of approximate machine unlearning for smooth strongly convex losses, showing that the optimal unlearning rate depends critica…

The paper introduces MINTS, a minimalist Bayesian framework that simplifies sequential decision-making by placing priors only on the optimum location, allowing for the incorporation of structural cons…

The paper introduces MeRa, a metric-space bias module, demonstrating that latent reasoning only improves spatial prediction when it is explicitly grounded in the underlying metric space.

DASH introduces a dual-branch distillation framework to effectively compress class-conditional diffusion models by independently supervising both score branches, significantly preserving guidance fide…

The paper proposes an algorithm for the extensive-form bandit problem that achieves $ ilde{O}(rac{ ext{total actions} imes ext{strategies} imes ext{trials}}{ ext{epsilon}})$ regret while satisfyi…

The paper introduces Sparse Memory-Efficient Training (SMET), a method that stabilizes and optimizes Dynamic Sparse Training (DST) for large language models, enabling stable and memory-efficient spars…

The paper systematically evaluates concept-based explainability in MLLMs, finding that forcing models to generate formal explanations degrades predictive accuracy, suggesting that explaining is genuin…