English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

GLARE: Adversarial Reward Estimation for Meeting Continuation and the MDFB Benchmark

Forum topic · 小凯 · 2026-09-15

Summary

A forum post introduces GLARE (Generative Learning via Adversarial Reward Estimation) alongside the Meeting Dynamic Forecasting Benchmark (MDFB), built from 2,207 real-world meetings and 24,794 future-facing queries. Given a transcript prefix and an active question, models generate multi-turn continuations in one call, evaluated on utility (progress toward the question) and human-likeness (plausible flow and role consistency) without requiring exact reproduction of the observed future. GLARE adapts adversarial imitation learning to conditional language generation: a discriminator ranks observed continuations above samples from the current actor, and its scores form a KL-regularized policy reward, with the discriminator retrained on negative samples from the evolving policy. In human evaluations GLARE achieves 0.66 utility win rate and 0.70 human-likeness win rate, surpassing SFT and SPIN but still below real human continuations. The authors also demonstrate MDFB as a social-reasoning arena for comparing general models, including closed-source systems, via reference-assisted judging. Paper: arXiv 2609.12165.

Overview

This forum post shares the paper GLARE: Generative Learning via Adversarial Reward Estimation (arXiv: 2609.12165), covering both the Meeting Dynamic Forecasting Benchmark (MDFB) and the proposed GLARE training method.

Field: Machine Learning Authors: Tenghao Huang, Zhaoxuan Tan, Muhao Chen, Jonathan May, Mengting Wan, Longqi Yang, Pei Zhou, Sihao Chen Posted: 2026-09-15

Key Points

  • Meeting Dynamic Forecasting Benchmark (MDFB): constructed from 2,207 real-world meetings and 24,794 future-facing queries. Meeting continuation requires tracking the agenda, speaker roles, participant intentions, and disagreement across long multi-party discussions.
  • Task setup: given a transcript prefix and an active question, a model generates a plausible multi-turn continuation in a single call.
  • Evaluation: two criteria — *utility* (progress toward the question) and *human-likeness* (plausible conversational flow and role consistency) — without requiring exact reproduction of the observed future.
  • GLARE method: an adaptation of adversarial imitation learning to conditional language generation. A discriminator ranks the observed continuation above samples from the current actor; its scores form a KL-regularized policy reward. The discriminator is retrained on negative samples from the current policy, so the reward landscape co-evolves with the actor.
  • Results: in human evaluations, GLARE achieves a 0.66 utility win rate and 0.70 human-likeness win rate, outperforming SFT and SPIN, though still below the observed human continuations.
  • Broader use: MDFB also serves as a social-reasoning arena for comparing general-purpose models, including closed-source systems, via reference-assisted judging — supporting both task-oriented learning and output-based evaluation of meeting behavior.
  • Links

  • Paper: https://arxiv.org/abs/2609.12165

Tags

#machine-learning#adversarial-imitation-learning#benchmarks#meetings#language-models#reinforcement-learning#arxiv#social-reasoning

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178634836