Large language models can't efficiently reason conditionally in a distribution-free setting.
problem Impossibility of conditional PAC-efficient reasoning in large language models.
method Proof of impossibility in a distribution-free setting for non-atomic input spaces.
result Any algorithm achieving conditional PAC efficiency must defer to the expert model with high probability.
RACER optimizes LLM-as-judge accuracy with dynamic reasoning selection.
problem Balancing reasoning accuracy with computational cost in LLM-as-judge settings.
method Formulates routing as a constrained distributionally robust optimization problem, accounting for distribution shift via KL-divergence uncertainty set.
result RACER achieves superior accuracy-cost trade-offs under distribution shift.
Abstractor enhances Transformers for relational reasoning, improving sample efficiency and performance.
problem Improving sample efficiency and performance in relational tasks.
method Introduces Abstractor module with relational cross-attention to enable explicit relational reasoning.
result Dramatic improvements in sample efficiency and performance on various relational tasks.
Fractured Sampling improves LLM reasoning efficiency by truncating CoT trajectories.
problem Efficiently scaling reasoning in large language models with limited tokens.
method Integrating truncated Chain-of-Thought (CoT) with Fractured Sampling across multiple dimensions.
result Fractured Sampling achieves superior accuracy-cost trade-offs compared to full CoT.
Kernel smoothing improves LLM reasoning efficiency.
problem Efficiently estimate value functions with limited samples for reinforcement learning.
method Kernelized advantage estimation using classical nonparametric statistics.
result Improved policy optimization with accurate value and gradient estimation.
Understanding and reasoning about physics is an important ability of intelligent agents. We develop the PHYRE benchmark for physical reasoning that contains a set of simple classical mechanics puzzles in a 2D physical environment. The benchmark is designed to encourage the development of learning algorithms that are sa…
Hydra boosts efficiency for long-context reasoning in resource-constrained settings.
problem Quadratic complexity of transformers limits long-context reasoning in resource-constrained systems.
method Hydra uses a modular architecture with adaptive routing between sparse global attention, mixture-of-experts, and dual memories.
result Hydra achieves significant throughput and accuracy improvements for long-context reasoning.
Paper introduces scalable neural architecture for solving NP-hard problems.
problem Solving NP-hard reasoning problems from natural inputs.
method Scalable neural architecture and loss function for discrete Graphical Models.
result Empirically shows efficient learning of NP-hard problems.
EORM boosts LLM accuracy with a lightweight, energy-based verifier.
problem Efficiently verifying mathematical reasoning in large language models.
method Energy-based framework to rank Chain-of-Thought solutions using simple outcome labels.
result EORM boosts LLM accuracy to 90.7% on GSM8k and 63.7% on MATH with only 55M parameters.
Transformers solve parity problems efficiently with step-by-step reasoning.
problem Training transformers to solve complex, recursive problems like parity.
method Training a one-layer transformer to solve k-parity, incorporating intermediate parities into the loss function, and using teacher forcing or augmented data. result Transformers can learn parity in one gradient update with intermediate supervision or self-consistency checks.
BRACE generates efficient counterfactual explanations by integrating causal reasoning.
problem Challenges in traditional counterfactual explanations, especially neglecting causal relationships.
method Backtracking counterfactuals with causal reasoning.
result Our method provides deeper insights into model outputs and is computationally efficient.
New method uses model comparison signals to improve LLM evaluation accuracy.
problem Limited benchmark sizes and model stochasticity in evaluating LLMs' mathematical reasoning.
method Combines standard labeled outcomes with model comparison signals to design a statistically efficient evaluation framework.
result Semiparametric estimator achieves the semiparametric efficiency bound and substantially improves ranking accuracy.
The paper studies how search and distillation improve reasoning in large language models.
problem Improving reasoning capabilities of large language models.
method Viewing chain-of-thought generation as a metastable Markov process, proving benefits of search and distillation.
result Search protocol rewards sparse edges, reducing the expected number of steps to reach different clusters.
Dynamic abstention improves LLM accuracy by selectively terminating unpromising reasoning.
problem LLMs waste compute on incorrect responses, leading to inefficiency.
method Formal reinforcement learning framework with abstention reward parameter.
result Dynamic abstention outperforms natural baselines in selective accuracy.
Early stopping methods reduce unnecessary reasoning steps in LLMs by monitoring uncertainty signals.
problem LLMs sometimes generate unnecessary reasoning steps, especially under uncertainty.
method Statistically principled early stopping methods that monitor uncertainty signals during generation.
result Uncertainty-aware early stopping improves efficiency and reliability in LLM reasoning, especially in math reasoning.
SFPO optimizes LLM reasoning by repositioning before updating, improving stability and efficiency.
problem Noisy gradients from low-quality rollouts cause instability and inefficient exploration in on-policy RL algorithms.
method Decomposes each step into three stages: a short fast trajectory, repositioning, and slow correction, preserving the objective and rollout process unchanged.
result SFPO consistently improves stability, reduces rollouts, and accelerates convergence, outperforming GRPO on math reasoning benchmarks.
ORCA calibrates LLMs for efficient, generalizable reasoning.
problem Miscalibration of large language models leading to inefficiencies.
method Online Reasoning Calibration (ORCA) using conformal prediction and test-time training.
result ORCA provides higher efficiency and generalization across different reasoning tasks.
Paper proves efficiency of MARL with transformers, addressing agent complexity.
problem Theoretical understanding of MARL with many agents and limited relational reasoning.
method Set Transformer for relational reasoning, model-free and model-based MARL algorithms.
result Provable efficiency of MARL algorithms, suboptimality gaps independent of number of agents.
FinReflectKG benchmarks financial QA by linking relevant context from a financial KG, improving model performance and efficiency.
problem Efficiently retrieving and navigating relevant financial information across diverse sources and years.
method A benchmark built on a temporally indexed financial KG, generating QA pairs via pattern-specific prompts and quality control, evaluating retrieval scenarios.
result KG-guided retrieval yields substantial gains in correctness and token utilization, improving model performance by 24%.
CausalEvolve improves efficiency and discovery in open-ended scientific tasks.
problem Lack of targeted guidance and knowledge organization in evolve-based agents.
method Causal scratchpad that identifies and reasons about guiding factors for evolution.
result Effective improvement in evolutionary efficiency and better solutions.
Hybrid framework injects TSLM insights into GRLM for robust time-series reasoning.
problem Lack of domain-specific knowledge in large language models for time-series reasoning.
method Hybrid knowledge-injection framework combining RLVR for efficient knowledge transfer.
result Consistently outperforms existing models by 7.9%-26.1% on multivariate time-series benchmarks.
MAC Net is a compositional attention network designed for Visual Question Answering. We propose a modified MAC net architecture for Natural Language Question Answering. Question Answering typically requires Language Understanding and multi-step Reasoning. MAC net's unique architecture - the separation between memory an…
LatFormer improves geometric reasoning by incorporating lattice symmetry priors in attention mechanisms.
problem State-of-the-art models struggle with geometric reasoning tasks in the ARC and LARC datasets.
method Introduced LatFormer, a model that uses lattice symmetry priors in attention masks.
result LatFormer requires 2 orders of magnitude fewer data than standard attention mechanisms.
This paper proposes an attention module augmented relational network called SARN(Sequential Attention Relational Network) that can carry out relational reasoning by extracting reference objects and making efficient pairing between objects. SARN greatly reduces the computational and memory requirements of the relational…
LASER compresses recursive model activations by exploiting their low-dimensional structure.
problem Understanding and optimizing the geometric structure of recursive reasoning trajectories.
method Dynamic low-rank basis tracking via matrix-free subspace tracking with a fidelity-triggered reset mechanism.
result Recursive activations occupy a linear, low-dimensional subspace that can be compressed efficiently.
CausAdv detects adversarial examples using causal reasoning.
problem Vulnerability of CNNs to adversarial perturbations.
method Causal framework based on counterfactual reasoning.
result Adversarial examples exhibit different CI distributions compared to clean samples.
Transformers improve logical reasoning on longer proofs but struggle with length.
problem Understanding systematic generalization in neural proof generation.
method Soft theorem-proving using Transformer models, evaluating logical consistency and inference accuracy.
result Transformers improve generalization with longer proofs but have difficulty with length.
EQD model improves domain-specific QA by 0.6% to 10.5%.
problem Challenges in domain-specific quantitative reasoning for LLMs.
method Two-step fine-tuning framework guided by a reward function.
result EQD outperforms state-of-the-art models and prompting strategies.
Develops models for temporally abstract reasoning and attention.
problem Temporal abstraction and attention in reinforcement learning.
method Defines affordances for options and develops partial option models.
result Identifies trade-offs between estimation and approximation error.
Tab-TRM uses recursive model for insurance pricing on tabular data.
problem Insurance pricing on tabular data.
method Adapts recursive latent reasoning to insurance modeling using a compact, parameter-efficient network.
result Improves insurance pricing accuracy using iterative refinement of latent tokens.
Improved RL for knowledge graph reasoning with entity types.
problem Challenges in path-based relational reasoning over knowledge graphs.
method Type-enhanced RL agent using GNN for neighborhood information.
result Outperforms state-of-the-art RL methods and discovers novel paths.
FedSight AI predicts federal funds rate using LLMs and multi-agent reasoning.
problem Predicting Federal Open Market Committee's decisions on federal funds rate.
method Multi-agent framework with large language models, structured and unstructured inputs, and CoD extension for efficient reasoning.
result Achieved 93.75% accuracy and 93.33% stability in predicting FOMC outcomes.
GraphToken encodes structured data for LLMs, improving graph reasoning tasks.
problem Efficiently encoding structured data for large language models.
method GraphToken learns an encoding function to extend prompts with explicit structured information.
result Significant improvements across node, edge, and graph-level tasks on the GraphQA benchmark.
Improved reasoning model by sampling from power distribution without additional training.
problem Efficiently sampling from a sharpened distribution to improve reasoning models.
method Entropy-Cut Metropolis-Hastings algorithm that identifies key decision points for resampling.
result The method consistently improves reasoning models across various datasets.
Attention mechanisms have been boosting the performance of deep learning models on a wide range of applications, ranging from speech understanding to program induction. However, despite experiments from psychology which suggest that attention plays an essential role in visual reasoning, the full potential of attention …
I summarize the recent work on market (in)efficiency, highlighting key elements why financial markets will never be made efficient. My approach is not by adding more empirical evidence, but giving plausible reasons as to where inefficiency arises and why it's not rational to arbitrage it away.
Improved neural networks for relational reasoning by projecting high-dimensional data to low-dimensional manifolds.
problem Out-of-distribution generalization in complex relational reasoning tasks.
method Neuroscience-inspired inductive-biased module projecting high-dimensional object representations to low-dimensional manifolds.
result Significantly better out-of-distribution generalization performance on relational reasoning tasks.
Knowledge graph reasoning, which aims at predicting the missing facts through reasoning with the observed facts, is critical to many applications. Such a problem has been widely explored by traditional logic rule-based approaches and recent knowledge graph embedding methods. A principled logic rule-based approach is th…
Survey examines distillation methods for large language models.
problem Efficiently compress large language models while preserving their capabilities.
method Knowledge Distillation and Dataset Distillation techniques.
result Integrating KD and DD can produce more effective and scalable compression strategies.
ERM uses energy-based selection to improve recursive reasoning.
problem Lack of principled inference mechanism in recursive models.
method Energy-guided Recursive Model (ERM) introduces Hopfield energies for trajectory selection.
result ERM achieves optimal solutions on various puzzles.
BASIS improves LLM reasoning by sharing batchwise rollout info, reducing MSE by 69%.
problem Improving large language model reasoning with limited rollouts and batch information.
method BASIS samples only one rollout per prompt but uses batch information to improve value function estimation.
result BASIS reduces MSE in value function estimation by 69% compared to REINFORCE++.
We present efficient differentiable implementations of second-order multi-hop reasoning using a large symbolic knowledge base (KB). We introduce a new operation which can be used to compositionally construct second-order multi-hop templates in a neural model, and evaluate a number of alternative implementations, with d…
Self-supervised method improves representation learning for better accuracy.
problem Improving representation learning without manual annotation.
method Proposes a novel self-supervised formulation of relational reasoning.
result Self-supervised relational reasoning outperforms state-of-the-art models by 14% in accuracy.
Paper closes neural-symbolic learning loop with grammar model and back-search algorithm.
problem Slow convergence in neural-symbolic learning due to error propagation issues.
method Introduces grammar model as symbolic prior and back-search algorithm for efficient error propagation.
result Significantly outperforms RL methods in performance, converging speed, and data efficiency.
CREAM models enable concept-grounded predictions and interpretability.
problem Designing models that can encode and extend prior knowledge about concept-concept and concept-task relationships.
method Proposes a flexible and efficient framework (CREAMs) that encodes arbitrary C−C and CoY relationships, incorporating a side-channel for incomplete concept sets. result CREAM models achieve competitive task performance while encouraging concept-grounded predictions, avoiding concept leakage and achieving black-box-level performance.
Sparse hypergraph neural networks improve reasoning in large knowledge graphs.
problem Reasoning about relationships in large, real-world domains using sparse and local inferences.
method Sparse and local hypergraph neural networks (SpaLoc) exploiting relational inferences that are usually local and sparse.
result State-of-the-art performance on real-world knowledge graph reasoning benchmarks.
Model learns brevity by exposing to easy problems, improving efficiency without explicit length penalties.
problem Excessive verbosity in step-by-step reasoning models trained with RLVR.
method Retaining and up-weighting moderately easy problems as implicit length regularizers.
result Model generates solutions that are, on average, nearly twice as short without explicit length penalties.
Researchers use DT to transfer policies from one environment to another using causal reasoning.
problem Adapting to changes in environmental dynamics in reinforcement learning.
method Applying causal counterfactual reasoning to Decision Transformer (DT) architecture for policy transfer.
result DT successfully transfers a learned policy to new environments while retaining most of the reward.