想不想知道，如何通过一场“团体赛”规则，治好AI“偏科”的毛病？又如何不花一分钱，只靠一个古老的“背包问题”，就让AI训练效率飙升40%？甚至，我们将看到一个天生“失明”的AI，如何在想象力测试中击败人类。本期节目，我们将从最新的几篇论文出发，揭示AI如何学会团队合作、聪明用钱，甚至掌握了“蒙眼思考”和“断舍离”的智慧。00:00:32 AI训练指南：如何让你的“学霸”不偏科？00:05:50 AI训练的“免费午餐”：如何不花一分钱，让模型变得更聪明？00:11:57 从“一指禅”到“组合拳”：AI学习的升级之路00:16:54 AI的“心灵盲区”：它看不见，却想得更明白？00:21:19 AI的记忆难题：如何优雅地“断舍离”？本期介绍的几篇论文：[LG] Polychromic Objectives for Reinforcement Learning [Stanford University] https://arxiv.org/abs/2509.25424 ---[LG] Knapsack RL: Unlocking Exploration of LLMs via Optimizing Budget Allocation [ByteDance Seed & The Chinese University of Hong Kong] https://arxiv.org/abs/2509.25849 ---[LG] Learning to Reason as Action Abstractions with Scalable Mid-Training RL [Apple] https://arxiv.org/abs/2509.25810 ---[LG] Artificial Phantasia: Evidence for Propositional Reasoning-Based Mental Imagery in Large Language Models [Northeastern University] https://arxiv.org/abs/2509.23108 ---[LG] Expected Attention: KV Cache Compression by Estimating Attention from Future Queries Distribution [NVIDIA & Sapienza University of Rome] https://arxiv.org/abs/2510.00636

Ask this episode

“What did the council decide on item 42?” — answered with the speaker and the moment.

Transcript

Related