Good Quant AI Papers

Curated top-conference research for quantitative finance and asset management.

Suggest a Paper

Scope

This is an evidence-bounded lower-bound catalog for 2024–2026 top-conference work with a direct contribution to quantitative investing, trading, portfolio construction, derivatives, or market-risk decisions. Every included entry has a verified venue record; an unverified preprint is not eligible. Main-conference, workshop, position, and affinity papers are labeled explicitly.

Coverage rows are deliberately conservative. A Pending venue-year means some official sources, side programs, or paper rosters remain unresolved; it is not a zero-eligible finding and should not be read as an exhaustive audit.

Included: asset allocation, alpha and factor modeling, market regimes, microstructure and execution, derivatives, market simulation, financial decision agents, and investment-linked alternative data.

Excluded: General banking, credit scoring, fraud detection, payments, accounting QA, regulatory technology, and generic financial NLP without a clear investment, trading, portfolio, or market-risk contribution.

The catalog stores original one-sentence editorial summaries and links—not paper PDFs or copied abstracts.

At a Glance

Metric	Value
Curated papers	154
Coverage units	33
Covered years	2024-2026
Conferences	11

How to Use

• Start with Browse by Year and Venue when you want a conference-specific list.
• Use Browse by Topic for stable generated topic pages over the controlled topic tags.
• Use data/papers.yaml for machine-readable paper metadata and data/coverage.yaml for the venue-year audit ledger.
• Read docs/metadata.md before changing fields or CONTRIBUTING.md before opening a pull request.

Coverage: 2024–2026

Venue	2026	2025	2024
ICML	20 papers · Pending	5 papers · Pending	3 papers · Pending
NeurIPS	0 papers · Pending	4 papers · Pending	14 papers · Complete
ICLR	1 paper · Pending	3 papers · Pending	3 papers · Complete
KDD	0 papers · Pending	7 papers · Pending	6 papers · Pending
AAAI	3 papers · Pending	0 papers · Pending	5 papers · Pending
IJCAI	5 papers · Pending	0 papers · Pending	14 papers · Pending
WWW	7 papers · Pending	0 papers · Pending	4 papers · Complete
WSDM	0 papers · Pending	0 papers · Pending	0 papers · Pending
SIGIR	0 papers · Pending	0 papers · Pending	0 papers · Pending
AISTATS	0 papers · Pending	1 paper · Pending	0 papers · No eligible papers
ACM ICAIF	0 papers · Pending	0 papers · Pending	49 papers · Pending

Browse by Year and Venue

• 2026 · ICML 2026 — 20 papers
• 2026 · ICLR 2026 — 1 paper
• 2026 · AAAI 2026 — 3 papers
• 2026 · IJCAI 2026 — 5 papers
• 2026 · WWW 2026 — 7 papers
• 2025 · ICML 2025 — 5 papers
• 2025 · NeurIPS 2025 — 4 papers
• 2025 · ICLR 2025 — 3 papers
• 2025 · KDD 2025 — 7 papers
• 2025 · AISTATS 2025 — 1 paper
• 2024 · ICML 2024 — 3 papers
• 2024 · NeurIPS 2024 — 14 papers
• 2024 · ICLR 2024 — 3 papers
• 2024 · KDD 2024 — 6 papers
• 2024 · AAAI 2024 — 5 papers
• 2024 · IJCAI 2024 — 14 papers
• 2024 · WWW 2024 — 4 papers
• 2024 · ACM ICAIF 2024 — 49 papers

Browse by Topic

Alpha Modeling · Alternative Data · Asset Allocation · Derivatives · Evaluation · Execution · Factor Investing · Financial Agents · Financial Forecasting · Market Microstructure · Market Regimes · Market Simulation · Portfolio Optimization · Risk Management · Synthetic Data

Data Files

• data/papers.yaml: canonical paper records with authors, field/topic tags, venue, year, track, URLs, identifiers, and original editorial summaries.
• data/coverage.yaml: venue-year audit status, official sources, checked tracks, pending tracks, and coverage notes.
• Generated Markdown lives in papers/ and topics/ and should not be edited by hand.

Research Watchlists

• docs/finance-journal-ai-seed-list.md: EDITH-local-corpus-backed seed list for technical AI, machine learning, and high-dimensional-data papers in JF, JFE, and RFS. Machine-readable metadata lives in data/finance_journal_ai_watchlist.yaml. It is intentionally separate from the canonical conference catalog.

Contributing

Have one candidate link? Suggest a Paper and the repository will extract base bibliographic metadata for maintainer review.

Contributions to sjsj0101/good-quant-ai-papers are welcome. Add or correct paper metadata in data/papers.yaml, update systematic coverage evidence in data/coverage.yaml when the venue-year audit state changes, provide an official venue source, and write original summary prose. Do not edit generated indexes by hand.

python3 scripts/validate.py
python3 scripts/render.py
python3 scripts/render.py --check

See CONTRIBUTING.md for the complete submission checklist.

Paper Index

ICML 2026

Main Conference (12)

Paper	Track	Focus	Assets / Frequency	Why it matters
A Linearly Convergent Proximal Subgradient Algorithm for Sparse Portfolio Optimization with Transaction Cost Xiaoting Yao, Na Zhang	Main Poster	Solves transaction-cost-aware sparse online portfolios through a difference-of-convex reformulation and proximal updates.	—	Adds a convergence-backed route to controlling both turnover costs and the number of active positions.
A Penalty Approach For Differentiation Through Black-box Quadratic Programming Solvers Yuxuan Linghu, Zhiyuan Liu, Qi Deng	Main Poster	Differentiates through black-box quadratic-program solvers using a smooth penalty surrogate and evaluates the method on multi-period portfolio optimization.	—	Connects solver-agnostic differentiation to a real portfolio decision while avoiding solver-specific KKT differentiation.
Adversarially Robust Control of Conditional Value-at-Risk via Kelly Conformal Inference Catherine Chen, Jingyan Shen, Xinyu Yang, Lihua Lei	Main Poster	Controls empirical conditional value-at-risk online with distribution-free guarantees under shifting or adversarial data.	—	Offers tail-risk safeguards for adaptive portfolio decisions without assuming a stationary return process.
Decision-focused Sparse Tangent Portfolio Optimization Haeun Jeon, Seunghoon Choi, Hyunglip Bae, Yongjae Lee, Woo Chang Kim	Main Poster	Trains a return model through differentiable asset selection and sparse tangency-portfolio reoptimization.	Equities	Targets risk-adjusted portfolio quality under a fixed holding budget rather than forecast accuracy in isolation.
Design Linear Constrained Neural Layers with Implicit Convex Optimization Junchi Yan, Liangliang Shi, Jiaxi Liu, Fangyuan Zhou, Wenzheng Pan, Zhongteng Gui, Yihui Tu	Main Poster	Enforces general linear constraints through implicit convex neural layers and evaluates the resulting constrained predictor on real portfolio allocation.	—	Lets neural systems honor hard portfolio-allocation constraints through a differentiable optimization layer.
Error Propagation in Dynamic Programming: From Stochastic Control to American Option Pricing Andrea Della Vecchia, Damir Filipovic	Main Poster	Analyzes how approximation errors accumulate through dynamic programs, including learned American-option valuation schemes.	Not Applicable	Clarifies when local model errors remain controlled in sequential pricing and exercise decisions.
Global Merger-Arbitrage Forecasting with Language Models Hinal Jajal, Michał Mucha, Charles Sweat, Chris Pulman, Charlie Flanagan, Peter Anderson	Main Poster	Predicts announced merger outcomes for global merger-arbitrage selection and scores deal-closing probabilities with P&L-weighted Brier loss.	—	Aligns language-model probability forecasts with the economic impact of concrete merger-arbitrage decisions across more than 400 deals.
Learning the ESG Geometry with Domain Aware Language Models Kunal Pradeep Pimparkhede, Chirayu Chaurasia, Jatin Roy, Mahesh Mohan Mohanachandran Radhamany	Main Poster	Learns domain-aware ESG representations and uses latent-space search for asset selection and downstream portfolio rebalancing.	—	Connects unstructured sustainability information to explicit security-selection and portfolio-rebalancing decisions.
Loss-aware distributionally robust optimization via trainable optimal transport ambiguity sets Jonas Ohnemus, Marta Fochesato, Riccardo Zuliani, John Lygeros	Main Poster	Learns loss-aware optimal-transport ambiguity sets end to end and tests whether they reduce conservatism in robust portfolio optimization.	—	Targets less conservative distributionally robust portfolio decisions by learning ambiguity sets for the downstream loss.
MarketSim: Simulating Stock Markets with Large-Scale Generative Agents Jinghua Piao, zhentao liu, Cheng Huang, Huang Jiarui, Songwei Li, Wang Ranran, Yong Li	Main Poster	Simulates stock-market dynamics through a large population of generative agents with heterogeneous behaviors.	Equities · Tick	Creates an experimental environment for studying emergent market outcomes and stress scenarios before live deployment.
Signature-Informed Transformer for Asset Allocation Yoontae Hwang, Stefan Zohren	Main Poster	Learns multi-asset allocations end to end using path-signature features and a downside-risk objective.	—	Connects market-path geometry directly to portfolio decisions instead of optimizing an intermediate forecast loss.
Tail Annealing for Heavy-Tailed Flow Matching Jean Pachebat	Main Poster	Evaluates extreme-tail and CVaR99 fidelity on controlled heavy-tailed benchmarks, then validates generation on real Fama–French equity-factor returns.	Equities	Combining CVaR99 fidelity with Fama–French validation makes the generator directly relevant to tail-sensitive market-risk scenario analysis.

Position Papers (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Position: Evaluating LLMs in Finance Requires Explicit Bias Consideration Yaxuan Kong, Hoyoung Lee, Yoontae Hwang, Alejandro Lopez-Lira, Bradford Levy, Dhagash Mehta, Qingsong Wen, CHANYEOL CHOI, Yongjae Lee, Stefan Zohren	Position Poster	Argues that finance LLM evaluation must explicitly control look-ahead, survivorship, and trading-cost biases.	Not Applicable	Point-in-time universes and realistic trading frictions determine whether reported forecasting and investment performance is valid.

Workshops (7)

Paper	Track	Focus	Assets / Frequency	Why it matters
Behavioral Proxy Conditioning for Financial Stress Scenario Generation with a Pretrained Diffusion Model Elena Kuular, Junsuk Choe	Workshop Poster Foundation Models for Structured Data	Conditions a pretrained diffusion model on behavioral proxies to generate financially meaningful stress scenarios.	—	Expands scenario design beyond historical replay while retaining interpretable links to stressed market behavior.
Forecast-to-Trade: Hierarchical Reinforcement Learning for Decision-Aware Financial Forecasting Zijie Zhao, Roy E. Welsch	Workshop Spotlight Forecasting as a New Frontier of Intelligence	Separates directional asset selection from constrained portfolio weighting in a hierarchical reinforcement-learning trader.	Equities · Daily	Evaluates forecasts through implementable rebalancing decisions that account for turnover, downside risk, and trading costs.
Leakage-Aware Benchmarking of LLM Forecasting: Real-Time Nowcasts as the Decision-Time Input for Macro Factor Ranking Mao Guan, Qian Chen	Workshop Poster Forecasting as a New Frontier of Intelligence	Benchmarks macro-conditioned equity-factor ranking using only data and nowcasts available at each historical decision time.	Equities	Separates genuine forecasting value from publication-lag leakage in investment backtests.
Learning to Trade Like an Expert: Cognitive Fine-Tuning for Stable Financial Reasoning in Language Models Yuchen Pan, Soung Chang Liew	Workshop Poster Foundations of Deep Generative Models: Understanding Memorization, Generalization, and Reasoning	Fine-tunes open language models on structured financial reasoning examples and evaluates transfer to chronological trading simulations.	—	Tests whether explicit decision reasoning produces more stable trading behavior across market regimes.
Mechanism-Inspired Aggregation for Multi-Agent Alpha Discovery: Optimizing Agent Distributions in Heterogeneous LLM Markets Ajitabh Kumar	Workshop Poster New Frontiers in Game-Theoretic Learning	Aggregates heterogeneous language-model agents by optimizing their population mix for multi-agent alpha discovery.	Equities · Daily	Treats agent diversity and weighting as part of the investment signal design rather than relying on simple voting.
Reflexivity as Prompt: Does Awareness of Self-Reinforcing Market Dynamics Improve LLMs as Financial Market Forecasters? Eugene W Park	Workshop Poster Forecasting as a New Frontier of Intelligence	Tests whether reflexivity-aware prompts improve market forecasts and evaluates an implied long/cash strategy across historical boom-bust episodes.	Multi Asset	Connects mechanism-aware forecasts to an economic strategy test through the implied strategy's Sharpe ratio.
TradeFM: A Generative Foundation Model for Trade-flow and Market Microstructure Srijan Sood, Maxime Kawawa-Beaudan, Daniel Borrajo, Manuela Veloso	Workshop Poster Foundation Models for Structured Data	Trains a generative trade-event model with scale-invariant features designed to transfer across equity markets.	Equities · Tick	Supports cross-market equity order-flow simulation and synthetic microstructure data without asset-specific tokenization.

ICLR 2026

Main Conference (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Trade in Minutes! Rationality-Driven Agentic System for Quantitative Financial Trading Zifan Song, Kaitao Song, Guosheng Hu, Ding Qi, Junyao Gao, Xiaohua Wang, Dongsheng Li, Cairong Zhao	Main Poster	Uses a rationality-driven multi-agent system to separate LLM strategy development from minute-level trading-bot deployment across stock and crypto markets.	Equities, Crypto · Intraday	Connects language-model strategy generation to executable trading policies with profitability, action-efficiency, and risk-control tests.

AAAI 2026

Main Conference (3)

Paper	Track	Focus	Assets / Frequency	Why it matters
FinRpt: Dataset, Evaluation System and LLM-based Multi-agent Framework for Equity Research Report Generation Song Jin, Shuqi Li, Shukun Zhang, Rui Yan	Main Not Specified AAAI Technical Track on Application Domains I	Introduces an equity-research-report benchmark, evaluation system, and LLM multi-agent framework using multiple financial data types.	Equities · Mixed	Moves LLM finance evaluation toward analyst-style research artifacts that can inform equity investment decisions and recommendations.
MARS: A Meta-Adaptive Reinforcement Learning Framework for Risk-Aware Multi-Agent Portfolio Management Jiayi Chen, Jing Li, Guiling Wang	Main Not Specified AAAI Technical Track on Machine Learning I	Combines heterogeneous portfolio-management agents with a safety critic and meta-adaptive controller to trade off return and downside risk.	Equities	Puts drawdown and volatility control directly into an RL allocation framework, which is closer to asset-management deployment constraints.
MetaTrader: Learning to Generalize RL Trading Policies Beyond Offline Data Haochen Yuan, Minting Pan, Yunbo Wang, Siyu Gao, Xiaokang Yang	Main Not Specified AAAI Technical Track on Machine Learning X	Learns offline reinforcement-learning trading policies for sequential portfolio optimization with transformation-based conservative temporal-difference updates.	Equities	Targets the offline-to-live generalization gap that often causes backtested RL trading policies to fail under new market conditions.

IJCAI 2026

Main Conference (5)

Paper	Track	Focus	Assets / Frequency	Why it matters
Beyond Isolated Investor: Predicting Startup Success via Roleplay-Based Collective Agents Zhongyang Liu, Haoyu Pei, Xiangyi Xiao, Xiaocong Du, Yihui Li, Suting Hong, Kunpeng Zhang, Haipeng Zhang	Main Oral Special Track on AI4Tech: AI Enabling Critical Technologies	Models venture-capital financing decisions as interactions among heterogeneous investor agents and compares their decisions with observed investor choices.	Not Applicable	Evaluates collective agents on a real private-market investment-screening decision rather than generic company prediction alone.
DiffLOB: Diffusion Models for Counterfactual Generation in Limit Order Books Zhuohan Wang, Carmine Ventre	Main Oral Special Track on AI4Tech: AI Enabling Critical Technologies	Generates regime-conditioned counterfactual limit-order-book trajectories and tests realism, intervention validity, and downstream predictive usefulness.	—	Supports market stress and scenario analysis with controllable synthetic microstructure paths.
Dual-Adversarial Dynamic Variational Asset Pricing with Adaptive Spatio-Temporal Feature Clustering for Portfolio Recommendation Yupeng Fang, Ruirui Liu, Xinyu Xia, Huichou Huang, Johannes Ruf, Qingyao Wu	Main Oral Main Track	Learns dynamic risk factors and exposures with variational, adversarial, and spatio-temporal components in a joint asset-pricing and portfolio-recommendation model.	Equities	Turns nonlinear risk-factor estimation into a directly evaluated stock recommendation decision under noisy returns and changing conditions.
TransAlpha: Lightweight Design Empowers Stock Return Forecasting Xiao Yang	Main Oral Main Track	Builds a lightweight cross-sectional Transformer for 15-minute A-share return signals and evaluates predictive power and portfolio profitability across broad equity-index segments.	Equities · Intraday	Links intraday return forecasts to portfolio profitability rather than stopping at statistical prediction accuracy.
Vector-Quantized Discrete Latent Factors Meet Financial Priors: Dynamic Cross-Sectional Stock Ranking Prediction for Portfolio Construction Namhyoung Kim, Jae Wook Song	Main Oral Main Track	Combines prior factors, vector-quantized latent factors, and regime-conditioned experts to rank CSI 300 and S&P 500 stocks and evaluate constructed portfolios.	Equities	Connects interpretable financial priors and latent market structure to security ranking and portfolio outcomes.

WWW 2026

Main Conference (6)

Paper	Track	Focus	Assets / Frequency	Why it matters
Analysis of CEX-DEX Arbitrage Opportunities with Hidden Markov Models Bence Ladóczki	Main Not Specified Economics, Online Markets and Human Computation	Uses hidden Markov models to characterize CEX-DEX price gaps and estimate arbitrage-trade counts and profits in Ethereum markets.	Crypto	Measures cross-venue crypto arbitrage opportunities and their economic value.
Financial Wind Tunnel: A Retrieval-Augmented Market Simulator Bokai Cao, Xueyuan Lin, Yiyan Qi, Chengjin Xu, Cehao Yang, Jian Guo	Main Not Specified Industry	Generates controllable cross-frequency market scenarios and uses them to stress-test and optimize downstream quantitative models under return and risk objectives.	Mixed	Provides decision-relevant synthetic scenarios for robustness testing and risk-aware model selection.
LiquidityPool: Game-Theoretic Analysis of Stakeholder Revenue in Ranking-Dependent DeFi Qinde Chen, Huawei Huang, Jian Zheng	Main Not Specified Economics, Online Markets and Human Computation	Analyzes how a ranking-dependent DeFi pooling mechanism changes participant revenue, concentration, welfare, and access on Ethereum.	Crypto	Provides a market-structure analysis of fund participation and revenue distribution in a crypto protocol.
Resisting Manipulative Bots in Meme Coin Copy Trading: A Multi-Agent Approach with Chain-of-Thought Reasoning Yichen Luo, Yebo Feng, Jiahua Xu, Yang Liu	Main Not Specified Security and Privacy	Develops a multi-agent defense for meme-coin copy trading and evaluates trading returns under manipulative bots and market frictions.	Crypto	Tests whether an automated trading decision can remain economically robust to adversarial manipulation.
Structure Over Scale: Diagnosing Liquidity Fragility in Concentrated-Liquidity AMMs Qiangqiang Liu, Runfa Jiang, Qian Huang, Frank Fan, Kunpeng Ren, Wei Cai	Main Not Specified Short Papers	Diagnoses structural sources of liquidity fragility in concentrated-liquidity automated market makers.	Crypto	Identifies crypto market-design conditions that affect liquidity resilience and market risk.
When Agents Trade: Live Multi-Market Trading Arena for LLM Agents Lingfei Qian, Xueqing Peng, Hanley Smith, Yi Han, Yueru He, Haohang Li, Yupeng Cao, Yangyang Yu, Guojun Xiong, Peng Lu, Yan Wang, Vincent Jim Zhang, Huan He, Alejandro Lopez-Lira, Jimin Huang, Jian-Yun Nie, Sophia Ananiadou	Main Not Specified Industry	Builds a live arena that continuously evaluates LLM agents making stock and cryptocurrency trades under distinct risk styles.	Equities, Crypto	Moves financial-agent evaluation from static answers to realized multi-market trading decisions and outcomes.

Workshops (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Learning-Based Optimization of Atomic Arbitrage in Decentralized Financial Systems Syahirul Faiz, Huned Materwala, Davor Svetinovic	Workshop Not Specified ZABAPAD 2026: 1st Workshop on Zero-knowledge Proof and Blockchain for WEB 4.0: Advancing the Post-quantum and Decentralized Era	Applies learning-based optimization to atomic arbitrage decisions in decentralized financial systems.	Crypto	Directly optimizes an executable DeFi trading strategy rather than merely describing protocol activity.

ICML 2025

Main Conference (5)

Paper	Track	Focus	Assets / Frequency	Why it matters
AlphaQCM: Alpha Discovery in Finance with Distributional Reinforcement Learning Zhoufan Zhu, Ke Zhu	Main Not Specified	Uses distributional reinforcement learning to search for complementary formulaic stock alphas under sparse, non-stationary rewards.	Equities	Automates the discovery of combined signals intended to improve stock selection across large equity universes.
Decision Making under the Exponential Family: Distributionally Robust Optimisation with Bayesian Ambiguity Sets Charita Dellaporta, Patrick O’Hara, Theodoros Damoulas	Main Not Specified	Builds posterior-informed ambiguity sets for distributionally robust decisions and evaluates them on portfolio optimization under model uncertainty.	—	Offers portfolio decisions that hedge estimation uncertainty while retaining tractable solve times and comparable robustness.
HyperIV: Real-time Implied Volatility Smoothing Yongxin Yang, Wenqi Chen, Chao Shu, Timothy Hospedales	Main Not Specified	Constructs arbitrage-free implied-volatility surfaces in real time from sparse option quotes and evaluates them across eight index options.	Derivatives	Supports faster, more stable derivative valuation and trading decisions when full volatility surfaces must be updated from limited market observations.
Latent Variable Estimation in Bayesian Black-Litterman Models Thomas Yuan-Lung Lin, Jerry Yao-Chieh Hu, Paul W. Chiou, Peter Lin	Main Not Specified	Treats Black–Litterman views and their uncertainty as latent variables learned from market data to generate stable portfolio weights.	Equities	Improves risk-adjusted allocation while reducing turnover relative to Markowitz and index baselines.
LOB-Bench: Benchmarking Generative AI for Finance - an Application to Limit Order Book Data Peer Nagy, Sascha Yves Frey, Kang Li, Bidipta Sarkar, Svitlana Vyetrenko, Stefan Zohren, Ani Calinescu, Jakob Nicolaus Foerster	Main Not Specified	Benchmarks generative limit-order-book streams with distributional, order-flow, price-response, and market-impact diagnostics.	Tick	Provides a decision-facing test suite for deciding whether synthetic microstructure data are credible enough for trading-model development.

NeurIPS 2025

Main Conference (4)

Paper	Track	Focus	Assets / Frequency	Why it matters
Online Portfolio Selection with ML Predictions Ziliang Zhang, Tianming Zhao, Albert Zomaya	Main Not Specified	Formalizes online portfolio selection with possibly wrong ML predictions and proposes an allocation rule with guarantees under perfect and adversarial forecasts.	Equities	Shows how modest predictive signals can be converted into wealth growth without discarding worst-case portfolio-selection guarantees.
OPHR: Mastering Volatility Trading with Multi-Agent Deep Reinforcement Learning Zeting Chen, Xinyu Cai, Molei Qin, Bo An	Main Not Specified	Uses a multi-agent reinforcement-learning architecture to time long/short volatility positions and route hedging strategies in cryptocurrency options.	Derivatives, Crypto	Converts volatility forecasts into explicit option-trading and hedging decisions evaluated by profit and risk-adjusted performance.
Robust Reinforcement Learning in Finance: Modeling Market Impact with Elliptic Uncertainty Sets Shaocong Ma, Heng Huang	Main Not Specified	Models market impact as asymmetric uncertainty and evaluates robust reinforcement-learning traders on single-asset and multi-asset trading tasks.	Multi Asset	Addresses the live-trading gap where a strategy's own orders move prices, improving robustness under larger trade volumes.
TwinMarket: A Scalable Behavioral and Social Simulation for Financial Markets Yuzhe YANG, Yifei Zhang, Minghao Wu, Kaidi Zhang, Yunmiao Zhang, Honghai Yu, Yan Hu, Benyou Wang	Main Not Specified	Simulates behavioral and social interactions among LLM agents in a stock-market environment to study bubbles, recessions, and other emergent market outcomes.	Equities	Provides a scalable experimental environment for stress-testing market dynamics that arise from interacting investor behaviors.

ICLR 2025

Main Conference (3)

Paper	Track	Focus	Assets / Frequency	Why it matters
An Online Learning Theory of Trading-Volume Maximization Tommaso Cesari, Roberto Colomboni	Main Not Specified	Studies how an online broker should set prices between two traders to maximize the number of mutually beneficial asset exchanges.	—	Gives regret guarantees for a microstructure-style execution objective where the broker observes trader responses rather than complete valuations.
MarS: a Financial Market Simulation Engine Powered by Generative Foundation Model Junjie Li, Yang Liu, Weiqing Liu, Shikai Fang, Lewen Wang, Chang XU, Jiang Bian	Main Not Specified	Builds an order-level generative foundation model and simulation engine for realistic, controllable financial-market trajectories.	Tick	Turns limit-order-book style data into a simulator for forecasting, risk detection, market-impact analysis, and trading-agent training.
Operator Deep Smoothing for Implied Volatility Ruben Wiedemann, Antoine (Jack) Jacquier, Lukas Gonon	Main Not Specified	Uses graph neural operators to map irregular option quotes into no-arbitrage-aware implied-volatility surfaces on intraday S&P 500 options data.	Derivatives · Intraday	Replaces repeated hand-calibrated option-surface fitting with a single learned operator for faster valuation, hedging, and risk workflows.

KDD 2025

Main Conference (7)

Paper	Track	Focus	Assets / Frequency	Why it matters
AlphaAgent: LLM-Driven Alpha Mining with Regularized Exploration to Counteract Alpha Decay Ziyi Tang, Zechuan Chen, Jiarui Yang, Jiayao Mai, Yongsen Zheng, Keze Wang, Jinrui Chen, Liang Lin	Main Not Specified Research Track	Builds an autonomous LLM-agent framework for generating, testing, and regularizing alpha factors to reduce crowding and decay.	Equities · Daily	Converts market hypotheses into candidate factor signals while explicitly penalizing repetitive, decay-prone alpha discovery.
CAMEF: Causal-Augmented Multi-Modality Event-Driven Financial Forecasting by Integrating Time Series Patterns and Salient Macroeconomic Announcements Yang Zhang, Wenbo Yang, Jun Wang, Qiang Ma, Jie Xiong	Main Not Specified Research Track	Integrates market time series with salient macroeconomic announcement text using causal and counterfactual augmentation for event-driven forecasting.	Multi Asset · Mixed	Links point-in-time macro news to tradable asset moves, which is central for event-risk monitoring and tactical allocation.
CryptoMixer: Fine-grained market information-aware MLP Networks for Individual Cryptocurrency Trading Prediction Tingsheng Feng, Zhihao Shen, Xi Zhao, Xiaoni Lu, Yuyang Zhou	Main Not Specified Research Track	Uses fine-grained market-information-aware MLP networks for individual cryptocurrency trading prediction.	Crypto	Focuses on asset-level crypto trading signals rather than generic market movement classification.
Enhancer: A Distribution-Aware Framework with Temporal-Relational Meta-Learning for Stock Prediction Weijun Chen, Shun Li, Heyuan Wang, Tengjiao Wang	Main Not Specified Research Track	Provides a model-agnostic temporal-relational meta-learning framework for adapting stock predictors under temporal and relational distribution shifts.	Equities · Daily	Addresses nonstationary market shifts that can degrade portfolio profitability, drawdown, and Sharpe ratios.
Multi-period Learning for Financial Time Series Forecasting Xu Zhang, Zhengang Huang, Yunzhi Wu, Xun Lu, Erpeng Qi, Yunkai Chen, Zhongya Xue, Qitong Wang, Peng Wang, Wei Wang	Main Not Specified Applied Data Science Track	Learns from short-, medium-, and long-horizon financial time-series signals for forecasting under changing market and policy conditions.	—	Helps investment and risk systems combine fast market changes with slower trend information in a single forecasting workflow.
Pre-training Time Series Models with Stock Data Customization Mengyu Wang, Tiejun Ma, Shay B. Cohen	Main Not Specified Research Track	Introduces stock-specific pretraining tasks for time-series transformers and evaluates downstream stock selection across multiple equity markets.	Equities · Daily	Improves return- and Sharpe-oriented stock selection by tailoring representation learning to market data structure.
Timing is important: Risk-aware Fund Allocation based on Time-Series Forecasting Fuyuan Lyu, Linfeng Du, Yunpeng Weng, Qiufang Ying, Zhiyan Xu, Wen Zou, Haolun Wu, Xiuqiang He, Xing Tang	Main Not Specified Applied Data Science Track	Uses time-series forecasting to drive timing-sensitive, risk-aware fund allocation decisions.	—	Maps forecasts into fund-allocation choices under risk constraints rather than optimizing prediction error alone.

AISTATS 2025

Main Conference (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Approximate Equivariance in Reinforcement Learning Jung Yeon Park, Sujay Bhatt, Sihan Zeng, Lawson L.S. Wong, Alec Koppel, Sumitra Ganesh, Robin Walters	Main Not Specified	Relaxes exact symmetry in reinforcement-learning policies and evaluates the resulting architecture on stock trading with real financial data.	Equities	Tests whether approximate market symmetries improve trading-policy performance and noise robustness without imposing unrealistic invariances.

ICML 2024

Main Conference (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Autonomous Sparse Mean-CVaR Portfolio Optimization Yizun Lin, Yangyu Zhang, Zhao-Rong Lai, Cheng Li	Main Poster	Builds a sparse mean-CVaR portfolio model that adapts its asset-selection mechanism as the investable universe changes.	—	Couples downside-risk control with scalable security selection instead of requiring a fixed asset pool.

Workshops (2)

Paper	Track	Focus	Assets / Frequency	Why it matters
Physics-Informed Neural Networks for Derivative-Constrained PDEs Kentaro Hoshisashi, Carolyn Phelan, Paolo Barucca	Workshop Poster AI for Science: Scaling in AI for Scientific Discovery	Extends physics-informed networks to PDEs with derivative constraints and benchmarks the method on option pricing and local-volatility calibration.	—	Enforces economically relevant derivative conditions while learning pricing and volatility surfaces from sparse observations.
You Shall Pass: Dealing with the Zero-Gradient Problem in Predict and Optimize for Convex Optimization Grigorii Veviurko, Wendelin Boehmer, Mathijs de Weerdt	Workshop Poster Differentiable Almost Everything Workshop	Repairs zero gradients in decision-focused learning and tests the resulting surrogate on a convex portfolio-optimization problem.	—	Makes it possible to train prediction models against downstream portfolio quality when the optimizer otherwise supplies no learning signal.

NeurIPS 2024

Main Conference (9)

Paper	Track	Focus	Assets / Frequency	Why it matters
A Globally Optimal Portfolio for m-Sparse Sharpe Ratio Maximization Yizun Lin, Zhao-Rong Lai, Cheng Li	Main Poster	Derives a globally optimal method for selecting an m-sparse portfolio that maximizes the Sharpe ratio.	—	Provides an exact risk-adjusted benchmark for portfolios constrained to hold only a small number of assets.
Action Gaps and Advantages in Continuous-Time Distributional Reinforcement Learning Harley Wiltzer, Marc G. Bellemare, David Meger, Patrick Shafto, Yash Jhaveri	Main Not Specified	Develops continuous-time distributional reinforcement learning and applies it to high-frequency American-option trading.	—	Models the distribution of trading outcomes when derivative decisions arrive faster than discrete-time methods naturally support.
Automated Efficient Estimation using Monte Carlo Efficient Influence Functions Raj Agrawal, Sam Witty, Andy Zane, Eli Bingham	Main Not Specified	Automates statistically efficient estimation with Monte Carlo influence functions and demonstrates it on optimal portfolio selection.	—	Quantifies portfolio objectives with principled uncertainty while reducing the need to derive custom estimators by hand.
Autoregressive Policy Optimization for Constrained Allocation Tasks David Winkel, Niklas Strauß, Maximilian Bernhard, Zongyue Li, Thomas Seidl, Matthias Schubert	Main Not Specified	Generates feasible allocation decisions autoregressively and evaluates the policy on constrained NASDAQ-100 portfolios.	—	Handles realistic portfolio constraints inside the learned decision process instead of repairing invalid weights afterward.
BPQP: A Differentiable Convex Optimization Framework for Efficient End-to-End Learning Jianming Pan, Zeqi Ye, Xiao Yang, Xu Yang, Weiqing Liu, Lewen Wang, Jiang Bian	Main Not Specified	Differentiates through convex programs efficiently and evaluates the framework on CSI-500 return-and-risk portfolio optimization.	—	Connects learned return signals directly to constrained portfolio decisions without an expensive generic solver backward pass.
FinBen: A Holistic Financial Benchmark for Large Language Models Qianqian Xie, Weiguang Han, Zhengyu Chen, Ruoyu Xiang, Xiao Zhang, Yueru He, Mengxi Xiao, Dong Li, Yongfu Dai, Duanyu Feng, Yijing Xu, Haoqiang Kang, Ziyan Kuang, Chenhan Yuan, Kailai Yang, Zheheng Luo, Tianlin Zhang, Zhiwei Liu, Guojun Xiong, Zhiyang Deng, Yuechen Jiang, Zhiyuan Yao, Haohang Li, Yangyang Yu, Gang Hu, Jiajia Huang, Xiao-Yang Liu, Alejandro Lopez-Lira, Benyou Wang, Yanzhao Lai, Hao Wang, Min Peng, Sophia Ananiadou, Jimin Huang	Main Poster Datasets and Benchmarks Track	Benchmarks language models across financial tasks that include stock trading, risk assessment, and investment decisions.	—	Makes model comparisons more credible by testing decision-facing finance capabilities in one reproducible suite.
FinCon: A Synthesized LLM Multi-Agent System with Conceptual Verbal Reinforcement for Enhanced Financial Decision Making Yangyang Yu, Zhiyuan Yao, Haohang Li, Zhiyang Deng, Yuechen Jiang, Yupeng Cao, Zhi Chen, Jordan W. Suchow, Zhenyu Cui, Rong Liu, Zhaozhuo Xu, Denghui Zhang, Koduvayur Subbalakshmi, Guojun Xiong, Yueru He, Jimin Huang, Dong Li, Qianqian Xie	Main Poster	Coordinates specialized language-model agents with verbal reinforcement to make sequential stock-investment decisions.	—	Tests whether explicit role specialization and reflection improve portfolio returns while controlling trading risk.
GLinSAT: The General Linear Satisfiability Neural Network Layer By Accelerated Gradient Descent Hongtai Zeng, Chao Yang, Yanzhen Zhou, Cheng Yang, Qinglai Guo	Main Not Specified	Enforces linear constraints in a differentiable layer and tests it on predictive equity portfolio allocation.	—	Keeps learned weights feasible while optimizing Sharpe ratio over a concentrated technology-stock universe.
Overcoming Brittleness in Pareto-Optimal Learning-Augmented Algorithms Alex Elenter, Spyros Angelopoulos, Christoph Dürr, Yanni Lefki	Main Not Specified	Robustifies learning-augmented online algorithms and applies them to one-way foreign-exchange trading decisions.	—	Balances the upside of exchange-rate predictions against worst-case protection when those predictions fail.

Workshops (4)

Paper	Track	Focus	Assets / Frequency	Why it matters
A Fully Analog Pipeline for Portfolio Optimization James S. Cummins, Natalia Berloff	Workshop Poster ML with New Compute Paradigms	Implements covariance estimation and minimum-variance portfolio construction as a fully analog computing pipeline.	—	Explores whether alternative hardware can accelerate the matrix operations behind efficient-frontier portfolios.
Do LLM Personas Dream of Bull Markets? Comparing Human and AI Investment Strategies Through the Lens of the Five-Factor Model Harris Borman, Anna Leontjeva, Luiz Pizzato, Max Kun Jiang, Dan Jermyn	Workshop Poster Workshop on Open-World Agents: Synnergizing Reasoning and Decision-Making in Open-World Environments (OWA-2024)	Compares human investors with personality-conditioned language-model personas in a simulated investment task.	—	Tests how risk appetite, impulsivity, and learning style change an AI agent's investment behavior before such personas are trusted with decisions.
InvestAlign: Align LLMs with Investor Decision-Making under Herd Behavior Huisheng Wang, Zhuoshi Pan, Hangjing Zhang, Mingxiao Liu, Yiqing Lin, H. Vicky Zhao	Workshop Poster Adaptive Foundation Models: Evolving AI for Personalized and Efficient Learning	Aligns language models with investor choices while explicitly modeling social-information cascades and herd behavior.	—	Tests whether an investment assistant can resist crowd-driven errors and preserve decision quality under social pressure.
Unlocking the Potential of Green Virtual Bidding Strategies : A Pathway to a Low-Carbon Electricity Market Aya Laajil, Laurent Barcelo, Frédérique M. Gagnon, Ghait Boukachab, Loubna Benabbou	Workshop Poster Tackling Climate Change with Machine Learning	Proposes using machine learning to explore green virtual-bidding strategies in financially settled day-ahead and real-time electricity markets.	—	Connects algorithmic bidding decisions to both market profitability and lower-carbon electricity-market operation.

Affinity Tracks (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
A Hybrid COMTE-LEFTIST Time-Series Explanation Method For a Time-series Classification Bitcoin Recommendation System lucas rabelo, Teresa Ludermir	Affinity Oral LatinX in AI	Combines COMTE and LEFTIST to explain a Bitcoin time-series classifier that turns one-minute closes into 30-minute sell-or-hold recommendations.	Crypto · Intraday	Makes a short-horizon crypto trading signal more interpretable by showing which time-series changes drive each recommendation.

ICLR 2024

Main Conference (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
New Insight of Variance reduce in Zero-Order Hard-Thresholding: Mitigating Gradient Error and Expansivity Contradictions Xinzhe Yuan, William de Vazelhes, Bin Gu, Huan Xiong	Main Poster	Reduces zeroth-order gradient variance in sparse hard-thresholding and demonstrates the method on portfolio optimization.	—	Makes black-box sparse allocation more practical when portfolio objectives can be evaluated but not differentiated.

Workshops (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
FinMem: A Performance-Enhanced LLM Trading Agent with Layered Memory and Character Design Haohang Li, Yangyang Yu, Zhi Chen, Yuechen Jiang, Yang Li, Denghui Zhang, Rong Liu, Jordan W. Suchow, Khaldoun Khashanah	Workshop Poster Workshop on Large Language Models for Agents	Organizes an autonomous trading agent's financial observations into layered memory before producing stock-investment decisions.	—	Shows how memory horizon and agent character affect cumulative returns on real-world equity data.

Affinity Tracks (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Logic-guided Deep Reinforcement Learning for Stock Trading Zhiming Li, Junzhe Jiang, Yushi Cao, Aixin CUI, Bozhi Wu, Bo Li, Yang Liu	Affinity Not Specified Tiny Papers @ ICLR 2024	Synthesizes a logic-guided hierarchy over reinforcement-learning subpolicies for robust stock trading.	—	Encodes market-trend knowledge directly in the strategy while improving return and limiting drawdown.

KDD 2024

Main Conference (4)

Paper	Track	Focus	Assets / Frequency	Why it matters
A Multimodal Foundation Agent for Financial Trading: Tool-Augmented, Diversified, and Generalist Wentao Zhang, Lingxuan Zhao, Haochong Xia, Shuo Sun, Jiaze Sun, Molei Qin, Xinyi Li, Yuqing Zhao, Yilei Zhao, Xinyu Cai, Longtao Zheng, Xinrun Wang, Bo An	Main Not Specified	Combines prices, news, charts, tools, memory, and reflection in an agent that trades stocks and cryptocurrencies.	—	Tests whether a generalist multimodal agent can turn diverse market evidence into profitable trading actions.
FreQuant: A Reinforcement-Learning based Adaptive Portfolio Optimization with Multi-frequency Decomposition Jihyeong Jeon, Jiwon Park, Chanhee Park, U Kang	Main Not Specified	Decomposes stock signals across frequencies before reinforcement learning adapts portfolio allocations to market shifts.	—	Uses persistent and transient return components to improve portfolio value and annualized performance.
MacroHFT: Memory Augmented Context-aware Reinforcement Learning On High Frequency Trading Chuqiao Zong, Chaojie Wang, Molei Qin, Lei Feng, Xinrun Wang, Bo An	Main Not Specified	Mixes context-specialized reinforcement-learning agents with memory to trade cryptocurrency at minute frequency.	—	Adapts high-frequency policies to changing trend and volatility regimes while preserving profitability.
Money Never Sleeps: Maximizing Liquidity Mining Yields in Decentralized Finance Wangze Ni, Yiwei Zhao, Weijie Sun, Lei Chen, Peng Cheng, Chen Jason Zhang, Xuemin Lin	Main Not Specified	Allocates cryptocurrency capital among decentralized-finance liquidity pools to maximize mining yield.	—	Formalizes yield farming as a dynamic allocation problem with executable pool-selection decisions.

Workshops (2)

Paper	Track	Focus	Assets / Frequency	Why it matters
FuNVol: A Multi-Asset Market Simulator using Functional Principal Components and Neural SDEs Vedant Choudhary, Sebastian Jaimungal, Maxime Bergeron	Workshop Oral Machine Learning in Finance	Simulates arbitrage-consistent multi-asset implied-volatility surfaces with functional components and neural stochastic differential equations.	—	Produces realistic cross-asset option scenarios whose quality is tested through delta-hedging profit-and-loss distributions.
Sourcing Investment Targets for Venture and Growth Capital Using Multivariate Time Series Transformer Lele Cao, Gustaf Halvardsson, Andrew McCornack, Vilhelm von Ehrenheim, Pawel Herman	Workshop Oral Machine Learning in Finance	Ranks prospective venture and growth-capital investments from multivariate company time series with a Transformer.	—	Turns proprietary operating histories into target-sourcing decisions and validates them with real investment portfolio simulations.

AAAI 2024

Main Conference (5)

Paper	Track	Focus	Assets / Frequency	Why it matters
CI-STHPAN: Pre-trained Attention Network for Stock Selection with Channel-Independent Spatio-Temporal Hypergraph Hongjie Xia, Huijie Ao, Long Li, Yu Liu, Sen Liu, Guangnan Ye, Hongfeng Chai	Main Not Specified	Pre-trains temporal and hypergraph representations before ranking stocks across dynamic market relationships.	—	Converts multivariate market structure into stock selections that improve investment returns and Sharpe ratios.
EarnHFT: Efficient Hierarchical Reinforcement Learning for High Frequency Trading Molei Qin, Shuo Sun, Wentao Zhang, Haochong Xia, Xinrun Wang, Bo An	Main Not Specified	Routes among specialized second-level reinforcement-learning traders to execute cryptocurrency strategies across market regimes.	—	Makes high-frequency policies trainable over long trajectories while improving simulated profitability and regime robustness.
ECHO-GL: Earnings Calls-Driven Heterogeneous Graph Learning for Stock Movement Prediction Mengpu Liu, Mengying Zhu, Xiuyuan Wang, Guofang Ma, Jianwei Yin, Xiaolin Zheng	Main Not Specified	Builds dynamic heterogeneous stock relations from earnings calls to forecast price movements over multiple horizons.	—	Turns earnings-call information into signals that improve both prediction accuracy and trading profitability.
Market-GAN: Adding Control to Financial Market Data Generation with Semantic Context Haochong Xia, Shuo Sun, Xinrun Wang, Bo An	Main Not Specified	Generates controllable financial-market time series conditioned on ticker, history state, and inferred market dynamics.	—	Provides context-aligned synthetic market scenarios for testing forecasts, risk models, and investment strategies.
MDGNN: Multi-Relational Dynamic Graph Neural Network for Comprehensive and Dynamic Stock Investment Prediction Hao Qian, Hongting Zhou, Qian Zhao, Hao Chen, Hongxiang Yao, Jingwei Wang, Ziqi Liu, Fei Yu, Zhiqiang Zhang, Jun Zhou	Main Not Specified	Encodes evolving multi-relational stock graphs with a Transformer to predict investment outcomes.	—	Captures changing links among stocks and related entities that static investment models miss.

IJCAI 2024

Main Conference (5)

Paper	Track	Focus	Assets / Frequency	Why it matters
Automatic De-Biased Temporal-Relational Modeling for Stock Investment Recommendation Weijun Chen, Shun Li, Xipu Yu, Heyuan Wang, Wei Chen, Tengjiao Wang	Main Not Specified	Debiases temporal stock relationships before producing investment recommendations aimed at excess returns.	—	Reduces relation-driven selection bias that can make graph-based equity recommendations unstable out of sample.
IMM: An Imitative Reinforcement Learning Approach with Predictive Representation Learning for Automatic Market Making Hui Niu, Siyuan Li, Jiahao Zheng, Zhouchi Lin, Bo An, Jian Li, Jian Guo	Main Not Specified	Combines imitation and predictive representation learning for automatic quoting at multiple price levels.	—	Learns market-making behavior from demonstrations while adapting quotes to evolving market states.
MacMic: Executing Iceberg Orders via Hierarchical Reinforcement Learning Hui Niu, Siyuan Li, Jian Li	Main Not Specified	Splits iceberg-order execution into hierarchical timing and sizing decisions conditioned on limit-order-book states.	—	Controls information leakage and execution cost for large hidden orders across diverse equities.
RSAP-DFM: Regime-Shifting Adaptive Posterior Dynamic Factor Model for Stock Returns Prediction Quanzhou Xiang, Zhan Chen, Qi Sun, Rujun Jiang	Main Not Specified	Adapts a dynamic stock-return factor model when latent market regimes shift.	—	Lets factor forecasts respond to structural changes rather than averaging incompatible return environments.
Trade When Opportunity Comes: Price Movement Forecasting via Locality-Aware Attention and Iterative Refinement Labeling Liang Zeng, Lei Wang, Hui Niu, Ruchen Zhang, Ling Wang, Jian Li	Main Not Specified	Uses locality-aware attention and refined labels to identify tradable price moves across stocks, ETFs, and cryptoassets.	—	Focuses the forecast on opportunities that translate into quantitative-investment performance across asset classes.

Workshops (9)

Paper	Track	Focus	Assets / Frequency	Why it matters
Comparing the Impact of Financial Knowledge Graphs from Financial Reports and Wikidata in Asset Recommendation Lubingzhi Guo, Javier Sanz-Cruzado, Richard McCreadie	Workshop Not Specified Recommender Systems in Finance	Compares report-derived and Wikidata knowledge graphs when ranking U.S. stocks for financial-asset recommendation and monthly portfolio returns.	Equities	Shows whether structured company knowledge can improve profitable security selection beyond price and news inputs alone.
Examining the Effect of News Context on Algorithmic Trading Surupendu Gangopadhyay, Prasenjit Majumder	Workshop Not Specified Joint Workshop of FinNLP and AgentScen	Trains a proximal-policy-optimization agent to combine news context with prices when choosing minute-level NIFTY 50 futures positions and quantities.	Derivatives · Intraday	Measures whether contextual news improves realized return, drawdown, volatility, Sharpe, and Sortino outcomes for an executable trading policy.
FAR-Trans: An Investment Dataset for Financial Asset Recommendation Javier Sanz-Cruzado, Nikolaos Droukas, Richard McCreadie	Workshop Not Specified Recommender Systems in Finance	Releases retail-investor transactions, security prices, and investor profiles with benchmarks that rank assets for future purchase and top-recommendation portfolio returns.	—	Provides a public decision-facing benchmark for comparing financial recommenders on both investor behavior and realized portfolio profitability.
GPT-Signal: Generative AI for Semi-automated Feature Engineering in the Alpha Research Process Yining Wang, Jinman Zhao, Yuri Lawryshyn	Workshop Not Specified Joint Workshop of FinNLP and AgentScen	Uses GPT-4 to generate formulaic alpha features and tests their ability to predict future returns and time buy and sell decisions for S&P 500 stocks.	Equities	Automates part of factor discovery while judging the generated signals by benchmark-relative investment performance.
LLM-Driven Knowledge Enhancement for Securities Index Prediction Zaiyuan Di, Jianting Chen, Yunxiao Yang, Ling Ding, Yang Xiang	Workshop Poster The First International OpenKG Workshop on Large Knowledge-enhanced Models	Combines language-model-derived market relations with daily index data in a heterogeneous graph to predict Shanghai Stock Exchange index trends and backtest the signals.	Equities · Daily	Tests whether inexpensive knowledge extraction can improve investable index-direction forecasts in a real market backtest.
Risk Propensity-specific Portfolio Recommendation via Self-supervised Learning Namhyoung Kim, Seung Eun Ock, Jae Wook Song	Workshop Not Specified Recommender Systems in Finance	Clusters Korean equities by volatility and tail-risk measures before constructing momentum portfolios matched to an investor's risk propensity.	Equities	Connects personalized recommendations to explicit return, volatility, Sharpe, drawdown, and cumulative-performance tradeoffs.
Sentiment trading with large language models Kemal Kirtac, Guido Germano	Workshop Poster AI4Research	Scores U.S. financial news with language models and forms value-weighted long, short, and self-financing long-short stock portfolios from the signals.	Equities · Daily	Demonstrates transaction-cost-aware portfolio gains and a 3.05 Sharpe ratio for the strongest language-model sentiment strategy.
Stock Recommendations for Individual Investors: A Temporal Graph Network Approach with Mean-variance Efficient Learning Youngbin Lee, Yejin Kim, Javier Sanz-Cruzado, Richard McCreadie, Yongjae Lee	Workshop Not Specified Recommender Systems in Finance	Models time-varying investor-stock interactions and mean-variance efficiency to recommend personalized equity portfolios with stronger return and Sharpe outcomes.	Equities	Aligns stock recommendations with portfolio diversification and profitability instead of optimizing recommendation relevance alone.
Wealth Guide at the FinLLM Challenge Task: A Sophisticated Language Model Solution for Financial Trading Decisions Sarmistha Das, R E Zera Marveen Lyngkhoi, Sriparna Saha, Alka Maurya	Workshop Not Specified Joint Workshop of FinNLP and AgentScen — FinLLM Shared Task	Maps stock and exchange-traded-fund prices plus financial news to explicit buy, sell, or hold actions in the FinLLM single-stock-trading task.	Equities	Demonstrates a language-model trading policy whose winning task result is evaluated with realized Sharpe ratio rather than text accuracy alone.

WWW 2024

Main Conference (3)

Paper	Track	Focus	Assets / Frequency	Why it matters
FinReport: Explainable Stock Earnings Forecasting via News Factor Analyzing Model Xiangyu Li, Xinjie Shen, Yawen Zeng, Xiaofen Xing, Jin Xu	Main Not Specified Industry Track	Combines news-factor analysis with explainable stock forecasts and evaluates the resulting signals in a transaction-cost-aware trading backtest.	—	Tests whether interpretable news-driven signals improve realized return, drawdown, and Sharpe outcomes rather than stopping at forecast accuracy.
Learning to Generate Explainable Stock Predictions using Self-Reflective Large Language Models Kelvin J. L. Koa, Yunshan Ma, Ritchie Ng, Tat-Seng Chua	Main Not Specified	Trains a self-reflective language model to explain next-day stock forecasts and validates the signals in portfolio construction.	—	Links text-based model explanations to realized portfolio outcomes rather than treating prediction accuracy as the final objective.
Reinforcement Learning with Maskable Stock Representation for Portfolio Management in Customizable Stock Pools Wentao Zhang, Yilei Zhao, Shuo Sun, Jie Ying, Yonggang Xie, Zitao Song, Xinrun Wang, Bo An	Main Not Specified	Learns a maskable stock representation so one reinforcement-learning allocator can operate across investor-selected equity universes.	—	Lets investors change the eligible stock pool without retraining a separate portfolio policy for every universe.

Workshops (1)

Paper	Track	Focus	Assets / Frequency	Why it matters
Measuring Arbitrage Losses and Profitability of AMM Liquidity Robin Fritsch, Andrea Canidio	Workshop Not Specified 3rd International Cryptoasset Analytics Workshop	Measures trading-fee income against arbitrage and loss-versus-rebalancing costs for automated-market-maker liquidity positions across pools and trading pairs.	—	Quantifies when liquidity provision is profitable and how block-time design changes the market risk borne by liquidity providers.

ACM ICAIF 2024

Main Conference (44)

Paper	Track	Focus	Assets / Frequency	Why it matters
A Financial Market Simulation Environment for Trading Agents Using Deep Reinforcement Learning Chris Mascioli, Anri Gu, Yongzhao Wang, Mithun Chakraborty, Michael P. Wellman	Main Not Specified	Builds an agent-based financial-market environment for training and evaluating deep reinforcement-learning traders.	—	Enables controlled experiments on strategy interaction and market outcomes before policies are exposed to live markets.
Adaptive and Explainable Margin Trading via Large Language Models on Portfolio Management Jingyi Gu, Junyi Ye, Guiling Wang, Wenpeng Yin	Main Not Specified	Uses language models to adapt and explain leveraged portfolio decisions under margin constraints.	—	Makes leverage selection responsive to market context while exposing the rationale behind portfolio-level risk taking.
Adaptive Risk-Based Control in Financial Trading Max M. Camilleri, Josef Bajada, Vincent Vella	Main Not Specified	Adapts a trading controller to changing risk conditions rather than applying a fixed exposure rule.	—	Lets a trading policy respond directly to evolving downside constraints during deployment.
Adversarial Inverse Reinforcement Learning for Market Making Juraj Zelman, Martin Stefanik, Moritz Weiss, Josef Teichmann	Main Not Specified	Recovers market-making behavior through adversarial inverse reinforcement learning.	—	Learns quoting objectives from observed behavior when the reward tradeoff among spread, inventory, and risk is unknown.
AI in Investment Analysis: LLMs for Equity Stock Ratings Kassiani Papasotiriou, Srijan Sood, Shayleen Reynolds, Tucker Balch	Main Poster	Evaluates language models as producers of equity ratings within an investment-analysis workflow.	—	Tests whether model-generated research opinions can support comparable and auditable stock-selection decisions.
ARL-Based Multi-Action Market Making with Hawkes Processes and Variable Volatility Ziyi Wang, Carmine Ventre, Maria Polukarov	Main Poster	Trains a multi-action market maker in a Hawkes-process environment with changing volatility.	—	Tests quoting policies against clustered order flow and volatility shifts that drive real inventory risk.
Augmenting Equity Factor Investing with Global Macro Regimes Dmitriy Nuriyev, Songyun Duan, Lingjie Yi	Main Poster	Conditions equity-factor strategies on learned global macroeconomic regimes.	—	Allows factor exposures to change when the macro environment alters expected returns and diversification benefits.
Autoregressive DRL with Learned Intrinsic Rewards for Portfolio Optimisation Magdalene Hui Qi Lim, Nixie S. Lesmana, Chi Seng Pun	Main Not Specified	Trains an autoregressive allocation policy with learned intrinsic rewards for sequential portfolio construction.	—	Shapes the reinforcement signal around portfolio quality when immediate market rewards are sparse or noisy.
Can GANs Learn the Stylized Facts of Financial Time Series? Sohyeon Kwon, Yongjae Lee	Main Not Specified	Tests whether generative adversarial networks reproduce the statistical regularities of financial time series.	—	Establishes whether synthetic paths preserve the market features needed for credible strategy and risk evaluation.
Cluster-driven Hierarchical Representation of Large Asset Universes for Optimal Portfolio Construction Nail Khelifa, Jérôme Allier, Mihai Cucuringu	Main Not Specified	Compresses large security universes into a hierarchical cluster representation for portfolio construction.	—	Reduces the dimensional burden of optimizing over many correlated assets while preserving cross-cluster diversification.
Contrastive Learning of Asset Embeddings from Financial Time Series Rian Dolphin, Barry Smyth, Ruihai Dong	Main Not Specified	Learns asset embeddings from return subwindows and evaluates their relationships in portfolio hedging and industry classification.	—	Supplies a nonlinear similarity measure for selecting diversifying hedge assets when correlation estimates are noisy.
Cross-Sector Market Regime Forecasting with LLM-Augmented News Analysis Timur Mudarisov, Radu Valentin State, Zsófia Kräussl, Alexander Yakubov, Tatiana Petrova	Main Poster	Combines cross-sector market data with language-model news representations to forecast regime changes.	—	Gives allocation and risk systems an early signal of sector-wide shifts that may not appear in prices alone.
Data-driven Derivative Hedging with Quadratic Variation Penalty Alessio Brini, Giacomo Domeniconi, Ali Fathi	Main Poster	Learns derivative hedges under a quadratic-variation penalty that discourages unstable trading paths.	—	Controls hedge variability alongside replication error, making learned strategies less sensitive to noisy rebalancing.
Deep Learning for Options Trading: An End-To-End Approach Wee Ling Tan, Stephen Roberts, Stefan Zohren	Main Poster	Learns option-trading decisions end to end from market inputs rather than optimizing a separate pricing forecast.	—	Aligns the model directly with derivative-strategy performance and trading frictions.
Denoising Diffusion Probabilistic Model for Realistic Financial Correlation Matrices Szymon Kubiak, Tillman Weyde, Oleksandr Galkin, Daniel Philps, Ram Gopal	Main Not Specified	Uses a denoising diffusion model to generate correlation matrices with financial dependence structure.	—	Supplies realistic covariance scenarios for portfolio-risk testing when historical matrices are noisy or data-limited.
Designing Expressive and Liquid Financial Options Markets via Linear Programming and Automated Market Making Xintong Wang, David M. Pennock, David M. Rothschild, Nikhil R. Devanur	Main Poster	Uses linear programming and automated market making to support expressive option contracts while preserving feasible prices.	—	Expands the range of hedgeable payoff structures without abandoning liquidity or internally consistent pricing.
Detecting Collective Liquidity Taking Distributions Andrei-Bogdan Balcau, Leandro Sánchez-Betancourt, Stefan Sarkadi, Carmine Ventre	Main Poster	Detects coordinated patterns in how market participants consume available liquidity.	—	Helps distinguish collective order-flow behavior that can change execution costs and short-horizon liquidity risk.
Dynamic Pricing in Securities Lending Market: Application in Revenue Optimization for an Agent Lender Portfolio Jing Xu, Yung-Cheng Hsu, William Biscarri	Main Poster	Dynamically prices securities loans to optimize revenue across an agent lender's inventory portfolio.	—	Coordinates lending fees across positions rather than maximizing each loan without regard to portfolio utilization.
Dynamic Reinforced Ensemble using Bayesian Optimization for Stock Trading Arishi Orra, Aryan Bhambu, Himanshu Choudhary, Manoj Thakur	Main Not Specified	Uses Bayesian optimization to tune a dynamic ensemble of reinforcement-learning stock traders.	—	Adapts strategy weights to changing performance rather than committing capital to one fixed trading policy.
DySTAGE: Dynamic Graph Representation Learning for Asset Pricing via Spatio-Temporal Attention and Graph Encodings Jingyi Gu, Junyi Ye, Ajim Uddin, Guiling Wang	Main Not Specified	Learns time-varying equity relationships with graph encodings and spatio-temporal attention for asset-pricing predictions.	—	Lets return signals incorporate changing cross-stock dependencies instead of relying on a static relation graph.
ECC Analyzer: Extracting Trading Signal from Earnings Conference Calls using Large Language Model for Stock Volatility Prediction Yupeng Cao, Zhi Chen, Qingyun Pei, Nathan Lee, K. P. Subbalakshmi, Papa Momar Ndiaye	Main Not Specified	Extracts trading signals from earnings-call language with a model trained to anticipate stock-volatility changes.	—	Converts management commentary into a volatility-aware signal that can inform event trading and risk sizing.
EX-DRL: Hedging Against Heavy Losses with EXtreme Distributional Reinforcement Learning Parvin Malekzadeh, Zissis Poulos, Jacky Chen, Zeyu Wang, Konstantinos N. Plataniotis	Main Not Specified	Models extreme loss tails inside distributional reinforcement learning for gamma-hedging options.	—	Improves the tail quantiles that determine VaR and CVaR for risk-sensitive derivative hedges.
Extracting Alpha from Financial Analyst Networks Dragos Gorduza, Yaxuan Kong, Xiaowen Dong, Stefan Zohren	Main Not Specified	Models relationships among financial analysts to derive stock-selection signals beyond standalone recommendations.	—	Treats the analyst network itself as alternative data for identifying differentiated equity views.
Fast Deep Hedging with Second-Order Optimization Konrad Mueller, Amira Akkari, Lukas Gonon, Ben Wood	Main Not Specified	Accelerates deep-hedging training with second-order optimization tailored to the hedging objective.	—	Reduces the computational burden of learning nonlinear derivative hedges under realistic risk criteria.
FinLlama: LLM-Based Financial Sentiment Analysis for Algorithmic Trading Giorgos Iacovides, Thanos Konstantinidis, Mingxue Xu, Danilo P. Mandic	Main Not Specified	Adapts a language model for financial sentiment signals evaluated in algorithmic trading.	—	Connects text classification to an executable trading rule instead of reporting sentiment accuracy alone.
FinVision: A Multi-Agent Framework for Stock Market Prediction Sorouralsadat Fatemi, Yuheng Hu	Main Poster	Coordinates language-model agents over news, candlestick charts, technical signals, and prior trade outcomes to predict stock moves.	—	Tests a multimodal research team whose forecasts feed concrete trading decisions and improve through post-trade reflection.
Hopfield networks for asset allocation Carlo Nicolini, Monisha Gopalan, Bruno Lepri, Jacopo Staiano	Main Not Specified	Formulates asset-allocation search through the energy dynamics of a Hopfield network.	—	Provides an alternative optimization mechanism for navigating combinatorial portfolio choices.
Machine Learning-based Relative Valuation of Municipal Bonds Preetha Saha, Jasmine Lyu, Dhruv Desai, Rishab Chauhan, Jerinsh Jeyapaulraj, Peter Chu, Philip Sommer, Dhagash Mehta	Main Poster	Learns comparable-value estimates for sparsely traded municipal bonds from cross-sectional bond attributes.	—	Helps fixed-income investors identify relative mispricing where transaction-based price discovery is thin.
Macroeconomic Conditioned Synthetic Financial Markets Alexander Michael Rusnak, Stéphane Daul	Main Not Specified	Generates synthetic financial markets conditional on macroeconomic states.	—	Supports scenario analysis in regimes that are scarce or absent from the historical return sample.
Market Making with Learned Beta Policies Yongzhao Wang, Rahul Savani, Anri Gu, Chris Mascioli, Theodore L. Turocy, Michael P. Wellman	Main Poster	Learns beta-distribution policies for quoting decisions in an agent-based market-making environment.	—	Gives liquidity providers a flexible policy class for balancing spread capture against inventory and adverse-selection risk.
Market-Making and Hedging with Market Impact using Deep Reinforcement Learning Jiayu Shi, Siu Hin Tang, Chao Zhou	Main Poster	Trains a joint market-making and hedging policy that accounts for the price impact of its own trades.	—	Integrates quoting, inventory control, and hedge execution instead of treating market impact as an afterthought.
Mixtures of Experts for Scaling up Neural Networks in Order Execution Kang Li, Mihai Cucuringu, Leandro Sánchez-Betancourt, Timon Willi	Main Poster	Scales neural execution policies through a mixture of specialized experts for heterogeneous order conditions.	—	Lets an execution system specialize across market states while retaining a shared policy architecture.
Neural Term Structure of Additive Process for Option Pricing Jimin Lin, Guixin Liu	Main Poster	Learns the maturity structure of an additive stochastic process for option valuation.	—	Provides a flexible way to fit option prices across expiries while retaining a coherent underlying process.
NeuralFactors: A Novel Factor Learning Approach to Generative Modeling of Equities Achintya Gopal	Main Not Specified	Learns latent equity factors inside a generative model of cross-sectional return behavior.	—	Offers a data-driven factor representation for simulating equity exposures and evaluating systematic strategies.
Numin: Weighted-Majority Ensembles for Intraday Trading Aniruddha Mukherjee, Rekha Singhal, Gautam Shroff	Main Poster	Aggregates intraday trading experts with a weighted-majority rule that adapts to their realized performance.	—	Diversifies short-horizon signal risk while shifting capital toward experts that remain effective online.
Optimizing Sequential Predictions for Order Execution: a Decision Focused Learning Approach Sunmin Kweon, Yonghwan Yim, Seungki Min	Main Poster	Trains sequential forecasts through the realized objective of an order-execution policy.	—	Rewards predictions that lower implementation cost rather than those that only minimize statistical error.
Quantum Generative Models of Mid-Price Movement in Limit Order Books Vanio Slavov Markov, Vladimir Rastunkov, Juan I Adame	Main Poster	Represents directional limit-order-book mid-price sequences with a parameterized quantum generative process.	—	Offers a new way to simulate short-horizon price dynamics for microstructure models and execution tests.
Reducing Return Volatility in Neural Network-Based Asset Allocation via Formal Verification and Certified Training Edward Stevinson, Alessio Lomuscio	Main Not Specified	Certifies and trains neural allocation rules to limit return volatility under specified input perturbations.	—	Adds formal robustness guarantees to portfolio policies whose weights can otherwise change unpredictably.
RiskMiner: Discovering Formulaic Alphas via Risk Seeking Monte Carlo Tree Search Tao Ren, Ruihan Zhou, Jinyang Jiang, Jiafeng Liang, Qinghao Wang, Yijie Peng	Main Poster	Searches symbolic factor expressions with a risk-seeking Monte Carlo tree policy to discover return signals.	—	Automates interpretable alpha discovery while focusing exploration on the most promising formula families.
Simulating Asset Prices using Conditional Time-Series GAN Riasat Ali Istiaque, Chi Seng Pun, Yuli Song	Main Poster	Generates conditional asset-price paths with a time-series adversarial model.	—	Expands the set of return scenarios available for portfolio stress tests and strategy development.
Stable Multilevel Deep Neural Networks for Option Pricing and xVAs Using Forward-Backward Stochastic Differential Equations Aadhithya Ashok Naarayan, Panos Parpas	Main Not Specified	Uses a stable multilevel neural solver for option prices and valuation adjustments represented by forward-backward stochastic equations.	—	Targets faster valuation of derivative exposures while controlling numerical instability across resolution levels.
Stock Recommendations for Individual Investors: A Temporal Graph Network Approach with Mean-Variance Efficient Sampling Youngbin Lee, Yejin Kim, Javier Sanz-Cruzado, Richard McCreadie, Yongjae Lee	Main Poster	Combines temporal investor-stock interactions with mean-variance sampling to personalize equity recommendations.	—	Aligns recommendation candidates with portfolio efficiency instead of optimizing relevance alone.
The Effect of Liquidity on the Spoofability of Financial Markets Anri Gu, Yongzhao Wang, Chris Mascioli, Mithun Chakraborty, Rahul Savani, Theodore L. Turocy, Michael P. Wellman	Main Not Specified	Studies how market liquidity changes the profitability and detectability of spoofing in a simulated trading environment.	—	Shows when apparent depth is most vulnerable to manipulation, informing execution and surveillance under different liquidity regimes.
Whack-a-mole Online Learning: Physics-Informed Neural Network for Intraday Implied Volatility Surface Kentaro Hoshisashi, Carolyn E. Phelan, Paolo Barucca	Main Poster	Fits intraday implied-volatility surfaces with a physics-informed network that incorporates derivative constraints during online updates.	—	Produces dynamically refreshed option surfaces while discouraging economically inconsistent shapes.

Workshops (5)

Paper	Track	Focus	Assets / Frequency	Why it matters
Hedging and Pricing Structured Products Featuring Multiple Underlying Assets Anil Sharma, Freeman Chen, Jaesun Noh, Julio DeJesus, Mario Schlener	Workshop Poster Simulation of Financial Markets and Economic Systems	Accelerates multi-asset autocallable pricing and learns a distributional-reinforcement-learning hedge for portfolios containing the notes.	—	Improves hedging profit-and-loss tails relative to delta-neutral and delta-gamma-neutral baselines while reducing pricing cost.
InvestorBench: A Benchmark for Financial Decision-Making Tasks with LLM-based Agent in Multimodal Market Environment Haohang Li, Yupeng Cao, Yangyang Yu, Shashidhar Reddy Javaji, Zhiyang Deng, Yueru He, Yuechen Jiang, Qianqian Xie, Jordan W. Suchow, K. P. Subbalakshmi, Zining Zhu, Jimin Huang	Workshop Not Specified Multimodal Financial Foundation Models	Benchmarks language-model agents on sequential buy, sell, and hold decisions across stocks, cryptocurrencies, and exchange-traded funds in multimodal market environments.	—	Compares decision quality with daily profit-and-loss and standard quantitative-finance metrics rather than language-only accuracy.
No Tick-Size Too Small: A General Method for Modelling Small Tick Limit Order Books Konark Jain, Jean-François Muzy, Jonathan Kochems, Emmanuel Bacry	Workshop Oral Simulation of Financial Markets and Economic Systems	Fits a Hawkes-process limit-order-book model that reproduces sparse and multi-level price dynamics across large-, medium-, and small-tick equities.	—	Provides a calibrated market simulator that preserves tick-size-specific liquidity and return behavior.
Supervised Autoencoder MLP for Financial Time Series Forecasting Bartosz Bieganowski, Robert Ślepaczuk	Workshop Poster Simulation of Financial Markets and Economic Systems	Trains supervised autoencoders with triple-barrier labels and noise augmentation to produce trading signals for equity-index, foreign-exchange, and cryptocurrency markets.	—	Evaluates the signals with realized strategy Sharpe and information ratios rather than forecast error alone.
TradingGPT: Elevating Financial Trading Performance in Multi-Modal Market Environments via LLM-Powered Multi-Agent Collaboration and Debate Yang Li, Yangyang Yu, Haohang Li, Zhiyang Deng, Yupeng Cao, Khaldoun Khashanah	Workshop Not Specified Multimodal Financial Foundation Models	Coordinates language-model agents that debate multimodal market evidence before selecting trading actions.	—	Evaluates whether structured multi-agent collaboration improves realized trading performance in a multimodal market environment.

License

Catalog metadata and original editorial prose are licensed under Creative Commons Attribution 4.0 International. Linked papers and third-party resources remain under their respective terms.