Unverified50% confidenceFactExact time
Tishby等人的信息瓶颈理论研究表明，网络训练后期会主动压缩与任务无关的信息表示
1
Sources
50%
Confidence
Long-term
Relevance
7/16/2026
First Seen
Sources
Related Claims
Unverified信息瓶颈理论由Tishby等人于2000年提出，主张深度网络在压缩输入信息的同时应保留与任务相关的最小充分统计量78% similarUnverifiedDuring cross-domain fine-tuning, an excessively large learning rate will destroy pre-trained weights, while an excessively small one will result in almost no learning, requiring a careful trade-off.69% similarUnverified端到端模型的核心优势在于避免模块间信息损失和误差累积，但代价是可解释性低、对训练数据分布敏感、难以应对长程推理任务67% similarUnverifiedNaïve post-training quantization causes individual weight errors to accumulate and amplify across multiple layers—a phenomenon known as compounding error—and at 2-bit or 1-bit precision the model structure collapses66% similarUnverified对于LLM推理场景，接口带宽瓶颈主要体现在初始加载速度上，而非持续推理性能66% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/530590API
curl https://kongchang.com/api/v1/knowledge/claims/530590MCP
get_claim(id=530590)