The workshop is taking place on 2nd November, 8am - 12pm NY (12pm - 4pm London) .
Click below to register.
IMPORTANT DATES
Paper submission: 15th September 2022 (Anywhere On Earth)
Author notification : 19th September 2022
Workshop: 2nd November 2022, 8am-12pm EST
Submission link: https://cmt3.research.microsoft.com/BAIF2022
Common benchmark problems and datasets have long been a driving force in AI. Shared benchmarks such as ImageNet, CIFAR, MuJoCo, Atari etc. have enabled researchers from around the world to tackle challenging problems and measure progress. The goal of this workshop is to spearhead an effort to create similar benchmarks for the subdiscipline of AI in Finance.
We invite submissions of benchmark datasets and environments for tasks in the financial domain, including (but not limited to):
Fraud detection, anti-money laundering, sanctions screening
Time series prediction
Market simulation
Optimal execution, market making, risk management
NLP/Information retrieval from financial documents
Credit risk scoring and decisioning
Customer experience - personalization, servicing
Automated policy compliance
Market mechanism design
Financial inclusion and coaching
We further invite submissions that discuss specific challenges to creating benchmarks for AI in finance (e.g. privacy concerns, regulatory constraints) and propose best practices and solutions (e.g. synthetic data). We also encourage position papers on best practices for designing benchmarks that can accelerate research and innovation in this field.
SUBMISSION GUIDELINES
Papers should be 4-6 pages long (not including references), and must be formatted according to the guidelines in Formatting section of ICAIF'22 submission instructions. There are no limits on length of supplementary materials, which could include dataset descriptions, code examples, data samples etc. Submissions can include previously published work as long as the focus for the workshop submission is on the benchmark and release of dataset/environment. The workshop is non-archival.
Submissions will be subject to single blind review, in order to allow authors to share datasets/environment easily during the review process and to make it available to the research community (e.g. through github) during/after the workshop. Links to the hosted dataset/environment would ideally be shared during the review process. If this is not possible, authors can share dataset samples/zipped code during the review, however we expect them to make it accessible (i.e. share links to hosted full dataset) before the workshop.
Benchmarks for AI in Finance website will feature all accepted submissions with links to hosted dataset/environment. Authors can choose hosting platform for datasets/environments. Authors should inform the workshop organizers if they foresee any difficulties in hosting their datasets/environment.