Topic
ML & AI Methods
Machine-learning methods applied to finance: deep learning, boosting, RL and new architectures.
- Papers featured
- 1,181
- Last 12 months
- 101
- Cited 100+
- 67
- Top venue
- SSRN
Featured per quarter
The last four quarters are highlighted.
Methods that come up
Click one to search for it.
Most cited
Featured papers in this topic with the most citations today.
- 5 Jun 202410,121cites
Mamba: Linear-Time Sequence Modeling with Selective State Spaces
Sequence Modeling: Mamba, a neural network architecture that doesn't use attention or MLP blocks, provides faster inference and better performance in language, audio, and genomics than Transformers.
Machine learning
- 22 May 20242,226cites
Octo: An Open-Source Generalist Robot Policy
Octo is a large transformer-based policy for robotic manipulation, trained on a vast dataset, that can be instructed via language or images and adapted to new domains.
Machine learningIn Robotics: Science and Systems Conference
- 5 Jun 20242,205cites
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality
The research identifies a link between state-space models and Transformers in deep learning, leading to the creation of a faster language modeling architecture, Mamba-2.
Machine learningIn International Conference on Machine Learning
- 24 Apr 20241,589cites
Mastering Diverse Domains through World Models
Algorithm Mastery: DreamerV3, a universal algorithm, excels in over 150 varied tasks, including diamond collection in Minecraft without human input, expanding the scope of reinforcement learning.
Machine learning
- 10 Jul 20241,271cites
SimPO: Simple Preference Optimization with a Reference-Free Reward
Simple Preference Optimization: SimPO improves reinforcement learning from human feedback by using the average log probability of a sequence as the implicit reward, enhancing training stability and computational efficiency.
Machine learningIn Neural Information Processing Systems
- 12 Jun 20241,269cites
Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction
The article discusses Visual AutoRegressive modeling (VAR), a new image learning method that outperforms diffusion transformers in terms of speed, image quality, and scalability.
Machine learningIn Neural Information Processing Systems
- 3 Jan 20241,032cites
PixArt-α: Fast Training of Diffusion Transformer for Photorealistic Text-to-Image Synthesis
PIXART-$\alpha$, a Transformer-based text-to-image model, generates high-quality images at a low cost, reducing CO2 emissions and offering a cost-effective solution for the AIGC community.
Machine learningIn International Conference on Learning Representations
- 8 May 2024817cites
Fourier Neural Operator with Learned Deformations for PDEs on General Geometries
The study introduces geo-FNO, a new framework for solving partial differential equations on any geometry, proving to be faster and more accurate than standard and machine learning-based solvers.
Machine learningIn J. Mach. Learn. Res.
- 3 Apr 2024737cites
LightGaussian: Unbounded 3D Gaussian Compression with 15x Reduction and 200+ FPS
LightGaussian is a new method that converts 3D Gaussians into a more compact format, enhancing efficiency in real-time neural rendering and reducing storage needs.
Machine learningIn Neural Information Processing SystemsFeatured 2×
- 13 Dec 2023628cites
Gated Linear Attention Transformers with Hardware-Efficient Training
Efficient Training of Gated Linear Attention Transformers: The research introduces a more hardware-efficient version of gated linear attention Transformers that performs well against other models, especially in training on longer sequences.
Machine learningIn International Conference on Machine Learning
- 22 May 2024513cites
CAT3D: Create Anything in 3D with Multi-View Diffusion Models
Multi-View Diffusion Models: CAT3D is a novel technique for generating 3D scenes from any number of images, surpassing existing methods in speed and efficiency.
Machine learningIn Neural Information Processing Systems
- 15 May 2024473cites
The Platonic Representation Hypothesis
The authors suggest that AI models, especially deep networks, are moving towards a common statistical model of reality, known as the platonic representation, and discuss its implications and limitations.
Machine learningIn International Conference on Machine Learning
Latest
- 6 Nov 20263fanfare
CosmosAlpha: Multi-Relational Graph Search with Pareto-Guided Exploration for LLM-Driven Alpha Mining
A graph-structured framework discovers alpha factors using an LLM, connecting factors by generation, similarity, and correlation while applying Pareto-guided search to balance quality, fertility and diversity.
arXiv
- 6 Nov 20263fanfare
LENS: Reliability-Aware Stock Prediction via LLM-Extracted Priors and News-Driven Semantic Graphs
A graph framework that weights extracted relations by reliability improves stock ranking on the S&P 500 by 16.4% in average recall and 22.6% in success rate.
arXiv
- 6 Nov 20264fanfare
Does It Actually Run? An Execution-Based Reproducibility Study of Published Machine-Learning Finance Research
Testing 459 quantitative-finance machine-learning papers' released code, the study finds that execution success rates fall steeply due to environment and dependency issues, not link rot alone.
SSRN
- 6 Nov 20263fanfare
Do AI Models Remember Indian Company Financials? Evidence on Look-Ahead Bias by Firm Size
Large language models recall Indian firm financials from within their training data with accuracy that declines sharply for smaller firms, creating serious look-ahead bias in LLM-based backtests.
SSRN
- 6 Nov 20263fanfare
Green Tastes, Beliefs, and Portfolios: Evidence from Synthetic Investors
Using language models as synthetic investors reveals institutional portfolios tilt greener despite weaker green tastes, and beliefs about green returns turn pessimistic after 2020.
SSRN
- 6 Nov 20264fanfare
Explaining Realized Returns: Roll’s R2 40 Years Later
Wall Street Journal market summaries explain 30-40% of daily and monthly returns out of sample when filtered for stated reasons, compared to under 10% for macro variables.
RePEc
- 6 Nov 20263fanfare
When Words Are Not Enough: Emotional Spillovers from Fed Chair Communication to Global Financial Markets
Facial and textual emotions from Fed Chair press conferences generate immediate effects on foreign exchange and equity returns, with non-additive cross-modal interactions.
RePEc
- 6 Nov 20263fanfare
Multilingual Textual Signals and Realized Volatility Forecasting: A Hybrid Transformer-HAR-X Approach Applied to the S&P 500 and the MASI
A hybrid FinBERT-HAR-X framework using multilingual news reduces S&P 500 volatility forecast error by 21% and improves realized Sharpe ratios out of sample.
RePEc
- 6 Nov 20263fanfare
The market's mirror: Revealing investor disagreement with LLMS
Using LLM personas, disagreement over news headlines predicts trading volume, with income and education disagreement most predictive and fundamentals generating least disagreement.
RePEc
- 6 Nov 20262fanfare
Artificial Intelligence and Financial Tail Risks
AI-scenario contributions to financial-sector tail risk vary substantially by country, with downside asymmetry in Europe and the UK but limited impact in North America.
RePEc
- 30 Oct 20264fanfare
Investor's Last Exam: Can AI Agents Do Equity Investment Without Seeing the Future?
Investor's Last Exam benchmarks AI agents on frozen financial databases, finding leading models score 68.1 overall yet only 52.5 on forecast accuracy.
arXiv
- 30 Oct 20263fanfare
Decorrelating the Advice Layer: A Calibrated Fragility Metric and Tested Mitigations for Synchronized Liquidation in Agentic Wealth Management
An agent-based model shows conversational AI wealth managers concentrate decisions into about two independent views, making markets 1.47 to 1.90 times more fragile than robo-advice.
arXiv
- 30 Oct 20263fanfare
Det-MASO: Diversity-Regularized Multi-Agent Sequential Optimization for Factor Generation
A multi-agent framework with a log-determinant penalty on agent correlations generates diversified alpha factors and improves information coefficient by 1.1 percent.
arXiv
- 30 Oct 20264fanfare
The AI Monoculture Paradox in Financial Markets
When traders adopt more precise but common AI signals, synchronized order flow can outweigh forecast gains, reducing collective profits despite higher individual precision, a paradox that taxes can help resolve.
SSRN
- 30 Oct 20263fanfare
Traders on GPT: AI and Cognitive Capital in Financial Markets
A model shows that cheaper AI substitutes for trader expertise and erodes cognitive skills, yet price informativeness remains stable until unexpected AI outages expose fragility in human expertise.
SSRN
- 30 Oct 20262fanfare
Leakage-Safe Multi-View Learning for Financial Transaction Fraud Detection under Temporal Shift
A multi-view learning framework for fraud detection integrates transaction, behavioral, temporal and graph features while avoiding look-ahead bias, with value depending strongly on the dataset.
SSRN
- 30 Oct 20262fanfare
Deep Learning Models in Central Bank Forecasting
Pretrained time-series foundation models match or outperform benchmarks on selected inflation and nowcasting targets but add limited gains for labor-market aggregates or irregular series.
RePEc
- 30 Oct 20262fanfare
Do AI Assistants Reinforce Investor Beliefs? Evidence from Human and AI Replies
AI-generated investment replies are systematically more bullish than human replies on Reddit; instruction-tuning amplifies bullishness beyond the base model's inherent bias.
RePEc
- 23 Oct 20260cites
Fake News, Real Crashes: Information Injection Attacks on LLM Financial Trading Agents
Fabricated news reaches 62% attack success against LLM trading agents, far exceeding prompt injection, with surface-form credibility cues mattering more than content accuracy.
arXiv
- 23 Oct 20260cites
Quant-Arena: Benchmarking LLM Agents for Quantitative Trading Research
A public benchmark tests whether large language models can complete end-to-end quantitative research with real trading constraints; no frontier model beats the equal-weight market in 2024.
arXiv
- 23 Oct 20260cites
Agentic AI in Automated Markets
Deep reinforcement learning simulates three agent archetypes in automated markets, finding herd-followers generate frequent short stress episodes while manipulators create persistent volatility.
arXiv
- 23 Oct 20260cites
Lineage-Aware Search Control for LLM-Based Evolutionary Factor Mining
LineageEvo guides large-language-model-based alpha factor search using lineage-conditioned variation and stratified allocation, promoting diverse and predictive factor discovery.
arXiv
- 23 Oct 20264fanfare
When Price Charts Rhyme: Historical Visual Retrieval for Stock Return Prediction
Vision models retrieve visually similar price charts from three decades of history; similar historical patterns predict future cross-sectional returns, especially for stocks with limits to arbitrage.
SSRN
- 23 Oct 20264fanfare
Artificial Intelligence and Firm Financing
As firms adopt AI, they hold less cash, shift from revolving to term credit, pay higher spreads, and rely more on non-bank lenders while expanding investment and employment.
SSRN
- 23 Oct 20263fanfare
Explainable Machine Learning for Basel and IFRS 9 Credit Risk: A Placebo-Controlled Test of Narrative Features
A placebo test reveals that features like central bank tone can raise credit model accuracy merely by indicating when loans were made, not by predicting defaults.
SSRN
- 23 Oct 20263fanfare
Where Machine Learning Helps: Risk Timing, Trend Filtering, and an Untouched 2025-26 Holdout
Five pre-registered hypotheses about machine learning for volatility, drawdowns, trend filtering, stock ranking, and news failed validation on a 2025-26 holdout despite passing development tests.
SSRN
- 23 Oct 20263fanfare
VQuant: An Agentic Large Language Model Architecture for Automated Quantitative Financial Analysis and Document Generation: Reasoning-Augmented Tool Orchestration, In-Loop Numerical Computation, and Multi-Format Reporting
A platform couples a language model with deterministic execution to turn natural-language requests into auditable reports where every figure is from live data or reproducible code, not model hallucination.
SSRN
- 23 Oct 20264fanfare
The Macroeconomic Effect of AI: Sizing the Software Engineering Channel
Using stock market data to measure AI's effect on software engineering productivity, the study estimates AI raised expected productivity by 32.6% and GDP by 3.6% to 6.5% through end-2025.
RePEc
- 23 Oct 20263fanfare
Large Language Models as Voting Mechanisms
The study proves that large language models function as voting rules and identifies millions of Condorcet cycles in ChatGPT investment recommendations, creating potential money-pump vulnerabilities.
RePEc
- 16 Oct 20260cites
What Do Foundation Models for Market Microstructure Learn? A Controlled Study with Synthetic Limit Order Books
Autoregressive transformers trained on limit-order-book message streams learn amortized Bayesian inference over market regimes, matching optimal prediction and enabling accurate parameter calibration.
arXiv