Menu
Sign In Search Podcasts Charts People & Topics Add Podcast API Pricing
Podcast Image

AI可可AI生活

AI前沿:从模型安全到推理加速

19 Apr 2025

Description

本期《TAI快报》介绍了五篇AI领域的突破性论文,涵盖模型安全、性能预测、模型设计、计算优化和推理增强: Antidistillation Sampling:提出反蒸馏抽样方法,通过“毒化”推理轨迹降低模型被蒸馏的风险,保护知识产权,同时维持模型性能。 Can Pre-training Indicators Reliably Predict Fine-tuning Outcomes of LLMs?:揭示传统困惑度预测微调性能的局限,提出Span Corruption困惑度和k-shot学习性能等新指标,提升模型选择效率。 It’s All Connected: A Journey Through Test-Time Memorization, Attentional Bias, Retention, and Online Optimization:通过Miras框架重新设计序列模型,提出Moneta等新模型,超越Transformer在长文本和推理任务中的表现。 Sleep-time Compute: Beyond Inference Scaling at Test-time:提出睡眠时计算范式,离线预处理上下文降低实时计算成本,减少5倍计算量并提升准确率。 Speculative Thinking: Enhancing Small-Model Reasoning with Large Model Guidance at Inference Time:提出推测性思考框架,利用大模型指导小模型推理,提升6-14%准确率并优化效率。完整推介:https://mp.weixin.qq.com/s/CF1EB3VugfcMlyKJbYpBFQ

Audio
Featured in this Episode

No persons identified in this episode.

Transcription

This episode hasn't been transcribed yet

Help us prioritize this episode for transcription by upvoting it.

0 upvotes
🗳️ Sign in to Upvote

Popular episodes get transcribed faster

Comments

There are no comments yet.

Please log in to write the first comment.