arXiv cs.CL· Manuel Prada-Corral, Yahya Emara, Timothy J. O'Donnell, Ryan Cotterell, Tim Vieira·· 10 小时前AI 评分27
PoolDPP:从语言模型中进行多样化采样的 DPP 模型
A Model of Diverse Sampling from Language Models
AI 导读
研究者提出 PoolDPP,将多样化生成形式化为从行列式点过程(DPP)中采样大小为 K 的完整字符串子集,似然编码质量、嵌入几何编码相似输出间的排斥。该方法先从可处理的提议分布抽取有限候选池,对候选做重要性加权后运行 K-DPP 采样器,并提供贪心 MAP 变体。
正文 · 原文
Abstract:Language models often produce high-quality individual samples but poor sets, with repeated samples clustering around the same semantic modes. We formalize diverse generation as sampling a size-$K$ subset of complete strings from a determinantal point process (DPP), where the likelihood encodes item quality, and the embedding geometry encodes repulsion between similar outputs. Exact inference over all strings is intractable, so we propose PoolDPP: draw a finite candidate pool from a tractable proposal, importance-weight the candidates toward the target quality, and run a $K$-DPP sampler on the induced pool kernel; a greedy MAP variant provides deterministic selection. We prove a finite-sample bound with an explicit constant: the KL divergence between the sampler's output and the target $K$-DPP decays as O(1/N) in the pool size. Empirically, PoolDPP extends the quality--diversity frontier on open-ended generation and the precision--recall frontier on ambiguous text-to-SQL (Ambrosia), and on mathematical reasoning (GSM8K), it raises diversity at about the same pass@$K$.
| Comments: | NeurIPS 2026 paper |
| Subjects: | Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Formal Languages and Automata Theory (cs.FL); Machine Learning (stat.ML) |
| Cite as: | arXiv:2611.25003 [cs.LG] |
| (or arXiv:2611.25003v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2611.25003 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Manuel Prada-Corral [view email]
[v1]
Wed, 18 Nov 2026 17:56:26 UTC (237 KB)
来源:arXiv cs.CL · arxiv.org