FinFM/27
Join PC Submit

AAAI 2027 Workshop FinHorizons Foundation

FinFM: Foundation Models
and Generative AI for Finance

A meeting point for the people building, evaluating, and deploying the next generation of financial AI.

Submit via OpenReview Join the Program Committee
  • February 22 or 23, 2027 Final day to be confirmed
  • Montréal, Canada Palais des congrès de Montréal
Explore AAAI 2027

Call for papers Submission deadline

Read the call

01The workshop

Beyond capability.
Toward reliability.

Foundation models and generative AI have moved from research prototypes to production systems in finance.

Yet the evidence behind these systems is fragile. Reported forecasting gains often fade under careful evaluation, and common practice remains exposed to look-ahead, survivorship and cost biases.

FinFM brings AI researchers, finance academics, practitioners and regulators together to establish how foundation models and generative AI for finance should be built, evaluated and validated.

Evidence before claims

Leakage-free benchmarks and reproducible, cost-aware evaluation.

Research meets practice

Open discussion across academia, industry, and financial institutions.

Resources for the community

A planned evaluation checklist and curated benchmark resources.

02Research directions

The questions
worth asking.

The topics of interest include, but are not limited to, the following.

01

Models & representations

Learning the language of financial markets.

  • Financial LLMs and continued pretraining
  • Time-series and order-book foundation models
  • Multimodal models for filings, news and prices
  • Generative scenario generation and stress testing
02

Agents & reasoning

From information to grounded decisions.

  • LLM agents for research, trading, compliance and advice
  • Retrieval-augmented and tool-augmented reasoning
  • Guardrails, hallucination control and safety
  • Deployment in regulated settings: small vs. frontier models
03

Evaluation & trustworthiness

Making progress measurable and reliable.

  • Leakage-free benchmarks and cost-aware evaluation
  • Uncertainty and calibration
  • Interpretability for model risk management
  • Systemic risk and model-risk validation

03Call for papers

Bring your research.
Move the field forward.

Submissions must be prepared using the AAAI-27 template and anonymized for double-blind review. We accept two kinds of paper.

Track 01

Research papers

Novel methods, theory, or experiments. Eligible for the Best Paper Award.

Track 02

Benchmark papers

New benchmarks, datasets, or evaluation protocols. Eligible for the Best Benchmark Award.

Long papers
Up to 7 pages of technical content, with unlimited pages for references and appendices
Short papers
Up to 4 pages of technical content, with unlimited pages for references and appendices
Supplementary material
Allowed, but reviewers are not required to review it
Awards
Determined by the quality of the work, regardless of page length

Official submission portal. Submit your paper through the FinFM venue on OpenReview. An OpenReview account is required.

Review process. Submissions are planned to undergo double-blind review on OpenReview, with at least two reviews and an organizer meta-review per paper. Please prepare an anonymized manuscript.

Presentations & proceedings. All accepted papers are planned to be presented as posters, with selected oral presentations and a benchmark spotlight. The workshop is non-archival by default; an opt-in online proceedings volume is planned. Final arrangements will be announced.

Important dates

  1. Submissions open 00:00 AoE
  2. Paper submission deadline 23:59 AoE
  3. Author notification
  4. Workshop22 or 23 Feb 2027 Final day to be confirmed

Submission times use Anywhere on Earth (AoE, UTC−12). The workshop day is still to be confirmed. See the tentative program.

Submit via OpenReview

04Program schedule

A program built
for exchange.

Tentative schedule

One full day on February 22 or 23, 2027, with keynote talks, oral presentations, poster sessions, an industry session, and a panel discussion. The final workshop date will be announced.

Morning · 09:00–12:30
– Opening remarks
– Keynote 1
– Oral session I3 papers · 15 minutes each
– Poster spotlight I30-second introductions
– Coffee break & poster session I
– Keynote 2
– Keynote 3
– Oral session II2 papers · 15 minutes each
Afternoon · 12:30–17:00
– Lunch break
– Keynote 4
– Keynote 5
– Industry session
– Coffee break & poster session II
– Panel discussionValidating Generative and Agentic Systems in Production
– Award ceremony & closing remarksBest Paper Award · Best Benchmark Award

All times are local to Montréal. Session times are provisional. Speakers, talk titles, and accepted papers will be announced after confirmation.

Keynotes include 25 minutes of presentation and 5 minutes of Q&A. Oral presentations include 10 minutes of presentation and 5 minutes of Q&A.

There are no criteria to attend beyond AAAI-27 registration. Anyone interested in foundation models and generative AI for finance is welcome. The number of attendees depends on the size of the room.

05Speakers

Speakers & panelists
to be announced.

Profiles of the keynote speakers, the industry session speaker, and the panelists will be posted here once they are confirmed.

Keynote speakers

Keynote 1

To be announced

Keynote 2

To be announced

Keynote 3

To be announced

Keynote 4

To be announced

Keynote 5

To be announced

Panel discussion

Validating Generative and Agentic Systems in Production

Panelist 1

To be announced

Panelist 2

To be announced

Panelist 3

To be announced

Panelist 4

To be announced

06Organizing committee

Different perspectives.
A shared ambition.

Connecting expertise in machine learning, quantitative finance, financial data science, and responsible AI.

Qingsong Wen

Squirrel Ai Learning
East China Normal University

AAAI Coordination

Building on a shared research community.

FinHorizons Foundation ICAIF 2025 · Rethinking Financial Time-Series

07Program committee

Expertise behind
every review.

Help shape the FinFM program. We are inviting reviewers who can give careful, constructive feedback on work at the meeting point of AI and finance.

Join the Program Committee
Who
Researchers and practitioners in machine learning, computational finance and financial NLP
Reviewing
Double-blind reviews on OpenReview. You tell us how many papers you can take when you apply.
When
Between the submission deadline on 20 November and author notification on 2 December 2026
Selection
The organizers review each applicant's publications before confirming appointments. Applying does not guarantee a place. Confirmed members will be listed here.

08Call for sponsors

Partner with
FinFM.

We are looking for sponsors who share our commitment to rigorous, reliable AI in finance. Sponsorship helps us recognize outstanding work and bring researchers, practitioners and regulators together. Every partnership is shaped around your goals.

Get in touch