Workshops

Large language models are increasingly deployed as autonomous agents in financial services, where they must reason, plan, explore, and adapt to dynamic environments. RL4LLM-Agents examines how reinforcement learning can move these systems beyond supervised fine-tuning, enabling them to optimize long-horizon objectives and learn from interaction in complex financial settings.

The workshop covers RLHF, RLAIF, GRPO, PPO, reward modelling, multi-agent market simulation, agentic retrieval, tool use, and sequential decision-making. It also emphasizes sample efficiency, safety, alignment, robustness, and the benchmarks and datasets needed to evaluate RL-trained financial agents under realistic, high-stakes conditions.

Responsible-AI research has advanced rapidly, but turning principles such as fairness, transparency, reliability, privacy, and accountability into production-grade financial systems remains difficult. RAIOps4Fin creates an interdisciplinary forum for the practical operationalization of responsible AI across the complete lifecycle of financial applications.

Contributions will emphasize empirical validation, deployment experience, and lessons from real-world systems. Topics include lifecycle monitoring and governance, model-risk management, validation, auditability, human oversight, regulatory implementation, responsible deployment of agentic AI, operational trade-offs, and energy-efficient and environmentally sustainable AI.

This workshop studies how artificial intelligence is reshaping market microstructure in both traditional and decentralized finance. It brings together work on trading, liquidity provision, price formation, and market design, with particular attention to methods that combine financial insight with machine learning, reinforcement learning, and agent-based modelling.

Topics include optimal execution, market making, trading under partial information, multi-agent models of strategic interaction, order-book dynamics, price discovery, decentralized exchanges, automated market makers, and on-chain liquidity. The programme also welcomes simulation, digital twins, generative market models, and data-driven approaches to fairness and regulation.

Women in AI and Finance supports early-career professionals navigating a field in which women continue to encounter barriers to participation and advancement. The workshop combines practical learning, mentorship, and community building, drawing on the experiences of women leaders working across artificial intelligence and financial services.

Sessions will examine AI's impact on finance, career opportunities and challenges, mentorship strategies, and the value of diverse professional networks. Interactive discussions and mentor–mentee networking are designed to help participants form lasting connections while contributing to a more inclusive and collaborative future for financial AI.

Financial AI is rapidly moving from prediction-oriented models toward foundation models, large language models, and autonomous agents supporting mission-critical services such as fraud detection, anti-money laundering, payments, and bank transfers.

The workshop treats security, privacy, and safety as interconnected requirements. It brings together researchers, practitioners, regulators, and financial institutions to discuss adversarial robustness, AI red teaming, privacy-preserving learning, governance, evaluation, and lessons from real-world deployments. Contributions may address secure LLMs and agents, financial-crime detection, federated learning, differential privacy, synthetic data, benchmarking, and operational case studies.

MFMB explores how multimodal and foundation models can combine tabular, time-series, text, image, and transactional data to address consequential banking problems such as credit scoring, fraud detection, pricing, and automated auditing.

The workshop focuses on the full deployment lifecycle: architectures and fusion strategies, trustworthy AI and regulation, efficient adaptation, privacy, evaluation under real operating conditions, causal reasoning, simulation, and synthetic data. It treats adoption as an end-to-end methodological problem rather than simply a model-architecture challenge. The programme will include keynote talks, contributed papers, and a panel discussion.

This workshop examines emerging AI applications across private equity, private credit, venture capital, real assets, and other alternative investments, bringing together researchers, practitioners, asset managers, allocators, data providers, and technology experts.

Its themes include agentic and multi-agent investment workflows, foundation models for diligence and document intelligence, fragmented and delayed private-market data, trustworthy AI for high-stakes illiquid decisions, and the impact of AI on portfolio companies and investment risk. Particular attention is given to evaluation with proprietary data, delayed outcomes, uncertain valuations, and limited comparables.

The workshop brings together AI researchers, financial machine-learning practitioners, quantum-ML specialists, and model-risk and governance experts to examine hybrid quantum-classical and quantum-inspired methods for financial services.

Its emphasis is rigorous evidence rather than unsupported claims of quantum advantage: reproducible benchmarks, strong classical baselines, decision relevance, model validation, and regulatory acceptability. Applications include trading, execution, hedging, credit, fraud, portfolio construction, risk forecasting, and scenario generation across tabular, time-series, graph, and multimodal data. The format will include invited talks, selected papers, posters, and a structured debate.

The third edition builds on the 2024 and 2025 ICAIF workshops and focuses on the next generation of agentic, tool-augmented, retrieval-grounded, and governance-aware generative AI systems for high-stakes financial workflows.

The workshop moves beyond isolated model capabilities toward reliable institutional systems, covering agent orchestration, harness engineering, workflow-level evaluation, red teaming, retrieval and tool use, governance, auditability, model-risk management, and human-agent collaboration. Applications span investment research, compliance, credit, risk, ESG, reporting, surveillance, and quantitative strategies, supported by new datasets, benchmarks, and infrastructure.

IAFM'26 addresses interpretability and alignment across financial decision-making, from credit scoring, fraud and AML detection, trading, and risk to large language models and agentic systems used in research, compliance, and client interaction.

The workshop bridges the established financial-XAI community with mechanistic interpretability and AI safety. Topics include post-hoc and concept-based explanations, interpretability of financial foundation models, generative-AI safety, faithful evaluation, regulatory compliance, model-risk governance, fairness, and production case studies. The programme combines peer-reviewed talks with a closing cross-sector panel involving research, industry, and regulatory perspectives.