LLMs

AnalysisBank: An Expert Analysis Pattern Library for Financial Report Generation
A library of expert analysis patterns, distilled from financial reports, that grounds LLM report generation in data-derived insights.
PDF Cite Code Project DOI arXiv
REPA: Reproducibility Evaluation via an Autonomous Pipeline Architecture
We introduce REPA, an LLM-powered pipeline that reproduces text classification experiments from paper descriptions through protocol extraction, input preparation, experiment generation, and evaluation.
PDF Cite Code Project OpenReview
Yuxi Xie Defends Her Ph.D. Thesis on Closed-Loop Scaling
Congratulations to Yuxi Xie on successfully defending her Ph.D. thesis, Closed Loop Scaling: Autonomous Improvement of LLM and VLM Reasoning.
8 May, 2026 WING, NUS
Yuxi Xie Defends Her Ph.D. Thesis on Closed-Loop Scaling
Lyle Ungar Visits WING to Give Invited NAII Seminar on Evaluating LLMs for Cross-Cultural Coaching and Therapy
Watch Lyle Ungar’s NUS AI Institute seminar on evaluating LLMs for culturally appropriate conversation in coaching and therapy.
30 Apr, 2026 11:00 — 12:30 COM1 Video Conference Room
Video
Lyle Ungar Visits WING to Give Invited NAII Seminar on Evaluating LLMs for Cross-Cultural Coaching and Therapy
KAHAN: Knowledge-Augmented Hierarchical Analysis and Narration for Financial Data Narration
A knowledge-augmented hierarchical framework that extracts insights from tabular data at entity, pairwise, group, and system levels for financial data narration.
PDF Cite Code Project DOI URL
TART – An Open-Source Tool-Augmented Framework for Explainable Table-Based Reasoning
A framework integrating LLMs with external tools to enhance explainability and accuracy in table-based reasoning tasks.
PDF Code Slides
TART – An Open-Source Tool-Augmented Framework for Explainable Table-Based Reasoning
Analytical Report Generation with LLMs
A challenging dataset for compositional reasoning and claim verification on scientific tables.
PDF Code
Analytical Report Generation with LLMs
SCITAB – A Benchmark for Scientific Table-Based Fact-Checking
A challenging dataset for compositional reasoning and claim verification on scientific tables.
PDF Code Slides
SCITAB – A Benchmark for Scientific Table-Based Fact-Checking