Tag: Classification

All the articles with the tag "Classification".

• Attention Retrieves, MLP Memorizes: Disentangling Trainable Components in the Transformer

  Published: 4 Jun, 2025 |  at  11:59 AM

  

  93.14 🤔

  本文通过冻结Transformer组件并提出MixiT模型，揭示了自注意力机制在检索和语言建模中的输入依赖性必要性，以及MLP层在记忆中的主导作用，强调了架构异质性对任务解决的重要性。

• Improving the Language Understanding Capabilities of Large Language Models Using Reinforcement Learning

  Published: 5 Jun, 2025 |  at  11:24 AM

  

  87.16 🤔

  本文通过将自然语言理解任务转化为强化学习问题，使用PPO算法微调中小规模LLMs，在GLUE和SuperGLUE基准上显著提升性能，超越监督微调和BERT-large，并展现出优于GPT-4o的零样本泛化能力。

• Unifying Attention Heads and Task Vectors via Hidden State Geometry in In-Context Learning

  Published: 31 May, 2025 |  at  11:19 AM

  

  86.54 🤔

  本文通过隐藏状态的几何特性（可分离性和对齐性）提出统一框架，揭示上下文学习（ICL）在分类任务中的两阶段机制——早期层通过PTH增强可分离性，后期层通过IH优化对齐性，并解释了任务向量的有效性。

• Label-efficient Single Photon Images Classification via Active Learning

  Published: 16 May, 2025 |  at  11:11 AM

  

  95.02 🤔

  This paper proposes an active learning framework for single-photon image classification that uses imaging condition-aware synthetic augmentation and a diversity-guided uncertainty-inconsistency sampling strategy to achieve high accuracy (97% on synthetic, 90.63% on real-world data) with significantly fewer labeled samples (1.5% and 8%, respectively) compared to baselines.

• Test-time Correlation Alignment

  Published: 8 May, 2025 |  at  12:21 AM

  

  93.31 🤔

  本文提出测试时相关性对齐（TCA）范式，通过构建伪源域相关性并应用线性变换对齐测试数据特征，显著提升测试时适应（TTA）性能，同时保持高效性和源域知识。