Research
On-device research index

arXiv research

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.

168,695 papers · 148 categories

Trend · papers per month

3467101134 · Jun 202019922001200920172026
48 results for Emergent Reasoning

This work explains crises in markets without external news using bounded rational agents.

problem Inability to model out-of-equilibrium dynamics in economic markets.
method Modeling bounded rational strategic reasoning in multi-agent market games.
result Bounded rational strategic reasoning can lead to endogenously emerging crises.

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.

Recent findings in multi-agent deep learning systems point towards the emergence of compositional languages. These claims are often made without exact analysis or testing of the language. In this work, we analyze the emergent language resulting from two different cooperative multi-agent game with more exact measures fo…

2020-01-23abs ↗pdf ↗

Discovering and exploiting the causal structure in the environment is a crucial challenge for intelligent agents. Here we explore whether causal reasoning can emerge via meta-reinforcement learning. We train a recurrent network with model-free reinforcement learning to solve a range of problems that each contain causal…

2019-01-23abs ↗pdf ↗

Bayesian RL enhances LLMs to reflectively explore and correct errors.

problem LLMs trained via RL lack reflective behaviors like rethinking and error correction.
method Bayesian RL framework that optimizes expected return under posterior distribution over Markov decision processes.
result BARL algorithm improves LLM performance in reasoning tasks.

A new method for math reasoning that allows for iterative correction.

problem Standard reasoning models commit to each token and cannot recover from early errors.
method Generative framework with latent thought vectors for iterative self-correction.
result 30 rethinking iterations surpass baselines with 15 times more parameters.

The study compares feed-forward and attention layers in language models.

problem Understanding the role of feed-forward and attention layers in language models.
method Empirical and theoretical analysis in a synthetic setting.
result Feed-forward layers learn simple distributional associations, while attention layers focus on in-context reasoning.

Transformers learn multi-step reasoning through gradient descent.

problem Understanding how transformers solve symbolic multi-step reasoning tasks.
method Theoretical analysis of gradient descent dynamics and multi-phase training.
result Trained one-layer transformers can solve both backward and forward reasoning tasks with generalization guarantees.

Survey on automating geometry problem solving with large models.

problem Automating geometric problem solving with spatial understanding and logical reasoning.
method Synthesizes GPS advancements through benchmark construction, parsing, and reasoning paradigms.
result Unified analytical paradigm and emerging opportunities identified.

XLVINs improve deep reinforcement learning by combining self-supervised learning and neural algorithmic reasoning.

problem Limitations of Value Iteration Networks (VINs) in deep reinforcement learning.
method Combining contrastive self-supervised learning, graph representation learning, and neural algorithmic reasoning.
result XLVINs match VIN-like models on discrete, fixed MDPs and significantly outperform model-free baselines.

AER dynamically adjusts entropy regularization for better LLM reinforcement learning.

problem Policy entropy collapse in RLVR training limits exploration and reasoning performance.
method Adaptive Entropy Regularization (AER) with difficulty-aware coefficient allocation, initial-anchored target entropy, and dynamic global coefficient adjustment.
result AER consistently outperforms baselines on mathematical reasoning benchmarks, improving both accuracy and exploration.

Bayesian framework explains diverse explanatory values.

problem Understanding and predicting human preferences for explanations.
method Developed a Bayesian account to integrate various explanatory values.
result Core values from psychology, statistics, and philosophy emerge from a common framework.

Look-ahead reasoning helps predict strategic user behavior on learning platforms.

problem Optimization criteria on learning platforms do not reflect users' priorities.
method Formalized level-k thinking and contrasted collective and selfish behavior.
result Coordination benefits users but does not offer higher-level reasoning advantages in the long run.

LLMs detect market patterns through causal reasoning, not just temporal association.

problem Detecting structural market patterns in financial data.
method Obfuscation testing using the WHO-WHOM-WHAT framework.
result LLMs achieve 71.5% detection rate of market patterns without temporal context.

Paper proves impossible for large language models to control hallucinations without sacrificing other properties.

problem Achieving truthful knowledge representation, semantic information conservation, and knowledge-constrained optimality simultaneously in large language models.
method Modeling inference as an auction of ideas, using mechanism design, proper scoring rules, and transformer architecture analysis.
result No LLM can simultaneously achieve all four essential properties without violating at least one.

Detecting the intention of drivers is an essential task in self-driving, necessary to anticipate sudden events like lane changes and stops. Turn signals and emergency flashers communicate such intentions, providing seconds of potentially critical reaction time. In this paper, we propose to detect these signals in video…

2019-05-03abs ↗pdf ↗

This work shows how approximate reward models can significantly improve inference-time scaling.

problem Improving the efficiency of inference for large language models.
method Identifying the Bellman error of approximate reward models and using Sequential Monte Carlo (SMC) for inference.
result Approximate reward models can reduce computational complexity from exponential to polynomial in TT.

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.

Geometric vector perceptrons improve protein structure learning.

problem Learning from protein structure with efficient and natural representations.
method Introducing geometric vector perceptrons to extend dense layers for Euclidean vectors, integrating geometric and relational reasoning.
result Improves model quality assessment and computational protein design over existing methods.

Benchmark evaluates financial misinformation detection models, revealing weaknesses without external context.

problem Detecting financial misinformation without external references.
method RFC Bench at paragraph level, two tasks: reference-free detection and comparison-based diagnosis.
result Performance improves with comparative context, revealing model weaknesses in reference-free settings.

Complexity is an interdisciplinary concept which, first of all, addresses the question of how order emerges out of randomness. For many reasons matrices provide a very practical and powerful tool in approaching and quantifying the related characteristics. Based on several natural complex dynamical systems, like the str…

2001-12-14abs ↗pdf ↗

CurveRL optimizes large model reasoning by reweighting prompts based on their rank and density.

problem Improving large language model reasoning through context reweighting.
method Formulated prompt reweighting as a functional derivative, proposing CurveRL based on quantile coordinate transform.
result CurveRL consistently outperforms existing methods across multiple benchmarks.

RLVR learning dynamics naturally create an implicit curriculum for transformers.

problem How rewards based on final outcomes help overcome the long-horizon barrier in reasoning models.
method Developed a theory of training dynamics for RLVR on transformers, using Fourier analysis on finite groups.
result Mixed-difficulty training induces an implicit curriculum that shapes the learning progression from easy to hard.

Proof confirms perfect representation in deep learning models.

problem Tackles the perfect Platonic Representation Hypothesis in deep learning models.
method Detailed proof using stochastic gradient descent (SGD) and analysis of global minima.
result SGD trains EDLNs to learn the same representation up to rotation, suggesting emergent entropic forces.

This paper improves learning complex functions with CoT supervision, reducing sample complexity.

problem Learning complex functions with multi-step reasoning.
method Develops a statistical theory linking CoT risk and end-to-end risk, using CoT information measure.
result CoT supervision can achieve significantly faster learning rates compared to standard E2E supervision.

CFRecs uses counterfactual reasoning to improve graph-based recommendations in real estate.

problem Improving model interpretability and actionable insights in graph-based recommender systems.
method A two-stage architecture combining GNN and Graph-VAE to propose minimal yet impactful changes in graph structure and node attributes.
result Demonstrates effectiveness in delivering actionable recommendations for home buyers and sellers.

Geometric framework explains and controls implicit bias in machine learning.

problem Understanding and controlling the selection of solutions in overparameterized models.
method Developed a theoretical and constructive framework based on geometric corrections induced by gradient noise and continuous symmetries of the loss.
result Computed the induced bias across various architectures and enabled inverse design to shape the bias.

Portfolio optimization emerged with the seminal paper of Markowitz (1952). The original mean-variance framework is appealing because it is very efficient from a computational point of view. However, it also has one well-established failing since it can lead to portfolios that are not optimal from a financial point of v…

2019-09-23abs ↗pdf ↗

We present an analysis of the price impact associated with trades effected by different financial firms. Using data from the Spanish Stock Market, we find a high degree of heterogeneity across different market members, both in the instantaneous impact functions and in the time-dependent market response to trades by ind…

2011-09-01abs ↗pdf ↗

LLMs learn new tasks from unstructured data, but it depends on word co-occurrence and positional information.

problem Understanding how LLMs can learn new tasks from unstructured data without explicit training.
method Examined the capabilities of LLMs trained on unstructured data, focusing on sequence model requirements and training data structure.
result Many ICL capabilities can emerge from word co-occurrence in unstructured data, but positional information is crucial for certain tasks.

Deep reinforcement learning algorithms have been successfully applied to a range of challenging control tasks. However, these methods typically struggle with achieving effective exploration and are extremely sensitive to the choice of hyperparameters. One reason is that most approaches use a noisy version of their oper…

2019-05-02abs ↗pdf ↗