Self-taught optimizer improves code generation using language models.
problem Improving code generation using language models.
method Recursive self-improvement of a scaffolding program that generates code.
result Improved scaffolding program generates programs with significantly better performance.
Self-improvement refines language models by verifying their own outputs.
problem Improving language models without external feedback.
method Formalizing self-improvement as sharpening, using the model itself as a verifier.
result RLHF-based self-improvement can outperform SFT-based methods.
The paper develops a theory for iterative self-improvement of models, proving conditions for better performance with easy-to-hard curricula.
problem Lack of theoretical foundation for iterative self-improvement in practical settings.
method Modeling self-improvement as maximum-likelihood fine-tuning on reward-filtered distributions and proving finite-sample guarantees.
result Explicit feedback loop and conditions for better performance with easy-to-hard curricula.
Classifier-based AI safety gates fail in self-improvement, even with advanced verification methods.
problem Maintaining reliable oversight of AI systems as they improve over iterations.
method Comprehensive empirical testing on neural controllers and MuJoCo benchmarks, using various classifiers and verification methods.
result Classifier-based safety gates fail in maintaining reliable oversight, even with advanced verification methods.
Theoretical limits on verifying self-improving systems without risking unbounded utility.
problem Formalizing and proving the limits of safety verification for self-improving systems.
method Developed dual conditions and used Holder's inequality, NP counting method, and Lipschitz bounds to establish impossibility and ceiling results.
result A classifier-based safety gate cannot simultaneously permit unbounded beneficial self-modification and bounded cumulative risk.
SRPO improves AI alignment with human preferences through self-improvement and task-independent optimization.
problem AI models trained with RLHF lack self-correction mechanisms and struggle with task generalization.
method SRPO formulates the preference learning problem as a min-max objective, optimizing a self-improvement policy and a generative policy in an adversarial fashion, making the solution task-independent.
result SRPO outperforms existing methods, achieving 90% AI Win-Rate on XSum and 56% on Arena-Hard prompts after a single revision.
RE enhances DL by learning model behavior, enabling iterative self-improvement.
problem Static data representations limit DL's potential for evolving models.
method RE uses multiple mappings of data through identical deep architectures, analyzing internal representations and performance signals.
result Models can gain insight from predecessors, leading to iterative self-improvement.
QuantAgent learns trading signals through self-improvement.
problem Building domain-specific knowledge for LLMs in quantitative investment.
method Two-layer loop approach: inner loop refines responses, outer loop tests and learns.
result QuantAgent approximates optimal trading behavior with provable efficiency.
SAIL-RevKL improves SAIL's convergence by regularizing the objective function.
problem Convergence of self-improving online LLM alignment algorithms.
method Proposed SAIL-RevKL, a regularized objective function to improve optimization landscape.
result Proved SAIL-RevKL satisfies the Polyak-Lojasiewicz (PL) condition with near-linear sample complexity.
SAIL improves online alignment of large language models with minimal feedback.
problem Offline RLHF methods often lead to sub-optimal performance due to fixed preference datasets.
method SAIL uses bilevel optimization and a single-level first-order method to iteratively refine model alignment.
result SAIL significantly improves alignment performance on open-sourced datasets with minimal computational overhead.
Active-GRPO improves molecular optimization by actively deciding when to imitate or self-improve.
problem Training robust and efficient molecular optimization models with large language models.
method Active-GRPO combines imitation and reinforcement learning, upgrading references and policies dynamically.
result Improves molecular optimization performance, achieving statistically significant gains.
SIBRE boosts reinforcement learning convergence by rewarding improvement over past performance.
problem Improving the rate of convergence in reinforcement learning.
method SIBRE is a reward shaping approach that rewards improvement over the agent's own past performance.
result SIBRE converges faster and more stably to the optimal policy compared to baseline RL algorithms.
In this article, we propose the Coopetititve Soft Gating Ensemble or CSGE for general machine learning tasks and interwoven systems. The goal of machine learning is to create models that generalize well for unknown datasets. Often, however, the problems are too complex to be solved with a single model, so several model…
LaTRO optimizes latent reasoning in LLMs without external reward.
problem Training LLMs to perform complex reasoning tasks.
method Formulates reasoning as latent distribution sampling and optimizes via variational approaches.
result LLMs improve reasoning and evaluation quality through self-improvement.
Model-free reinforcement learning based methods such as Proximal Policy Optimization, or Q-learning typically require thousands of interactions with the environment to approximate the optimum controller which may not always be feasible in robotics due to safety and time consumption. Model-based methods such as PILCO or…
RL pipeline simplifies knot diagrams, including very hard unknots.
problem Simplifying complex knot diagrams, especially very hard unknots.
method Reinforcement learning for move proposals and heuristic navigation of Reidemeister moves.
result Trained agent simplifies diagrams, including a 4 1 # 9 10 4_1\#9_{10} 4 1 # 9 10 link to a three-step unknotting process. Study compares DSPy teleprompter algorithms for aligning LLM evaluations with human annotations.
problem Aligning LLM evaluation metrics with human annotations.
method Comparative analysis of five teleprompter algorithms within the DSPy framework.
result Certain teleprompters outperform others in detecting hallucinations.
Creating efficient deep neural networks involves repetitive manual optimization of the topology and the hyperparameters. This human intervention significantly inhibits the process. Recent publications propose various Neural Architecture Search (NAS) algorithms that automate this work. We have applied a customized NAS a…
Neural model accelerates SDDP for stochastic optimization.
problem Exponential complexity of SDDP limits its applicability to low-dimensional problems.
method Trainable neural model maps problem instances to a low-dimensional piecewise linear value function.
result ν-SDDP significantly reduces problem solving cost without sacrificing solution quality.
Paper addresses online alignment of large language models under uncertain preference feedback.
problem Online alignment of large language models with misspecified preference feedback.
method Formulates an oracle-robust objective as a worst-case optimization problem for log-linear policies, and develops projected stochastic composite updates.
result Shows that the robust objective admits an exact closed-form decomposition and achieves O ~ ( ε − 2 ) \widetilde{O}(\varepsilon^{-2}) O ( ε − 2 ) oracle complexity. New algorithms reduce dueling bandits' regret with neural networks and efficient exploration.
problem Optimizing dueling bandits with neural networks for better performance.
method Combines shallow exploration strategies with neural networks for utility approximation, using iterative self-improvement and spectral analysis to reduce network width.
result Achieves sublinear regret of O ~ ( d ∑ t = 1 T σ t 2 + d T ) \widetilde{\mathcal{O}}(d\sqrt{\sum_{t=1}^{T} σ_t^2} + \sqrt{dT}) O ( d ∑ t = 1 T σ t 2 + d T ) . This paper presents a new meta-modeling framework to employ deep reinforcement learning (DRL) to generate mechanical constitutive models for interfaces. The constitutive models are conceptualized as information flow in directed graphs. The process of writing constitutive models are simplified as a sequence of forming g…
Transcribing content from structural images, e.g., writing notes from music scores, is a challenging task as not only the content objects should be recognized, but the internal structure should also be preserved. Existing image recognition methods mainly work on images with simple content (e.g., text lines with charact…
Sequence probability predicts correctness in LLMs, but not for repeated prompts
problem Predicting correctness in large language models
method Quantifying sequence probability and correctness across different levels
result Higher sequence probability often predicts correctness across prompt-answer pairs
Optimal self-distillation improves generative models' velocity risk and mode recovery.
problem Improving generative models' velocity risk and mode recovery.
method Proved optimal self-distillation for rectified flow via linear probing, derived mixing coefficient, and provided validation tuning.
result Optimal self-distillation improves velocity risk and mode recovery.
Self-play fine-tuning improves diffusion models for text-to-image generation.
problem Plateauing performance of diffusion models after data saturation.
method Self-play fine-tuning (SPIN-Diffusion) using competition among model versions.
result Significantly improved model performance and human preference alignment.
SUPE combines unlabeled data with RL to efficiently explore tasks.
problem Efficient exploration in reinforcement learning with sparse rewards.
method Extract low-level skills using VAE, pseudo-label unlabeled data, and use as off-policy data for online RL.
result SUPE outperforms prior methods across 42 long-horizon tasks.
Improves many-shot learning by optimizing and generating influential examples.
problem Limited performance in many-shot in-context learning (ICL).
method Iterative optimization and generation of influential examples.
result Significant improvements across various tasks using BRIDGE.
Meta-learning adapts models for unseen tasks across AI, robotics, and NLP.
problem Adapting models to unseen tasks efficiently and accurately.
method Black-box, metric-based, layered, and Bayesian approaches.
result Meta-learning enhances model generalization and adaptation to unseen tasks.
The paper tackles strategic behavior in decision-making with counterfactual explanations.
problem Finding optimal counterfactual explanations and policies in a strategic setting.
method NP-hard problem, greedy algorithm, submodularity, randomized algorithm, matroid constraint.
result Optimal counterfactual explanations and policies increase utility.
Overview of integrable systems with symmetries, focusing on toric and semitoric systems.
problem Classifying and understanding integrable systems with symmetries.
method Using decorated polygons and controlled bifurcations in one-parameter families of systems.
result Construction of explicit semitoric systems with prescribed invariants.
New method to derive integrable systems from existing Lax systems.
problem Deriving new integrable systems from existing ones.
method Systematic method of deriving new integrable systems from a given one.
result Examples of new integrable systems derived, including the dispersionless Hirota equation, the general heavenly equation, and the web equations.
Learning to control linear systems is statistically hard, especially for underactuated systems.
problem Statistical difficulty of learning to control linear systems, especially underactuated ones.
method Utilized minimax lower bounds and structural assumptions to prove learning complexity can be exponential.
result Learning complexity can be at most exponential with the controllability index of the system.
Discrete-time systems can be characterized by simple flat coordinates and their shifts.
problem Characterizing flatness of discrete-time systems.
method Developed a map from flat coordinates and their shifts to system state and input, fulfilling system equations identically.
result Derived necessary conditions for a system to be flat, without requiring differential geometry methods.
The paper explores when linear system identification is hard or easy, especially for under-actuated systems.
problem Statistical hardness of learning linear systems, especially under-actuated or under-excited systems.
method Using tools from minimax theory and recent statistical tools for finite sample analysis of system identification.
result The controllability index of linear systems affects the sample complexity of identification, making some systems hard to learn.
This paper improves system identification by reducing sample complexity for high-dimensional linear dynamical systems.
problem High sample complexity for learning partially observed linear dynamical systems in high dimensions.
method Introduces an ℓ 1 \ell_1 ℓ 1 -regularized estimation method that reduces sample complexity from linear to logarithmic with system dimension. result Markov parameters can be learned with logarithmic number of samples relative to system dimension, improving sample complexity.
In integrable hydrodynamic systems, coordinates exist where generators and symmetries are simple.
problem Existence of Riemannian invariants for integrable systems of hydrodynamic type.
method Finding coordinates where the generator and all symmetries are diagonal.
result In integrable hydrodynamic systems, there exist coordinates where the generator and all symmetries are diagonal.
This paper studies nonholonomic constraints in Hamiltonian systems, deriving equations and theorems.
problem Analyzing nonholonomic constraints in Hamiltonian systems.
method Deriving distributional RCH systems, geometric constraint conditions, and Hamilton-Jacobi theorems.
result Derives precise geometric constraint conditions and Hamilton-Jacobi theorems for nonholonomic systems.
New method models unknown systems with hidden parameters using neural networks.
problem Modeling unknown dynamical systems with hidden parameters.
method Training a deep neural network (DNN) model using trajectory data of the unknown system.
result DNN model accurately predicts unknown dynamical systems with new initial conditions.
Study absolute equivalence for Pfaffian systems, applying to control systems.
problem Absolute equivalence of Pfaffian systems with specific independence conditions.
method Structural results for Pfaffian systems of corank 3, applied to control systems.
result Dynamic feedback linearization of control systems with 2 inputs.
Estimates input from output of nonlinear systems using ANN.
problem Estimating unknown compositional input from system output.
method Artificial Neural Networks (ANNs) for nonlinear system inversion.
result ANNs can compete with optimal bounds for linear systems and demonstrate promising results for nonlinear systems.
Solves selecting the best optimizing system problems.
problem Selecting the best system among contenders with unknown performance.
method Adaptive algorithms integrating stochastic gradient descent and sequential elimination.
result Exponential rates of convergence to zero for false selection probability.
This paper considers control systems defined on Lie algebroids. After deriving basic controllability tests for general control systems, we specialize our discussion to the class of mechanical control systems on Lie algebroids. This class of systems includes mechanical systems subject to holonomic and nonholonomic const…
Superintegrable systems on curved manifolds found to have Hessian structures.
problem Characterizing superintegrable systems on curved manifolds.
method Identifying and computing Hessian coordinates for superintegrable systems.
result Examples of superintegrable systems in 2D and 3D have natural Hessian coordinates.
Systemic risk refers to the risk that the financial system is susceptible to failures due to the characteristics of the system itself. The tremendous cost of systemic risk requires the design and implementation of tools for the efficient macroprudential regulation of financial institutions. The current paper proposes a…
Researchers solve boundary and scattering rigidity problems for magnetic systems.
problem Recovering magnetic systems from boundary or scattering data.
method Reduced to magnetic systems and applied results from [DPSU07].
result Recovering M P \mathcal{MP} M P -system up to a gauge. The paper develops methods to derive mixed superposition rules for Lie systems and applies them to various physical systems.
problem Finding general solutions for Lie systems.
method Develops mixed superposition rules for Lie systems with imprimitive Lie algebras and semidirect sums.
result Extends coalgebra method to Lie systems of partial differential equations.
Paper develops reduction theory for controlled Lagrangian systems with symmetry and momentum map.
problem Reduction of controlled Lagrangian systems with symmetry and momentum map.
method Using Legendre transformation and Euler-Lagrange vector field, the paper extends symmetric reduction theory.
result Established regular reduction theory for RCL systems with symmetry and momentum map.