Unverified50% confidenceFactExact time
2014年前后以CTC和注意力机制为核心的端到端ASR模型开始崛起，将识别过程统一为序列到序列映射问题
1
Sources
50%
Confidence
Long-term
Relevance
8/5/2026
First Seen
Sources
Qwen-Audio-3.0-ASR-Flash发布：医疗术语召回率达95%的语音识别新标杆
Related Claims
UnverifiedThe attention mechanism was first proposed by Bahdanau et al. in 2014 to address the information bottleneck problem of long-range dependencies in sequence-to-sequence models.61% similarUnverifiedCTGAN引入模式特定归一化通过高斯混合模型分解连续变量的多峰结构，并通过训练时条件采样机制缓解类别不均衡59% similarUnverifiedThe VRAM occupied by the context window is calculated separately from the model weights, and KV cache VRAM usage grows linearly with context length.59% similarVerifiedThe self-attention mechanism enables models to simultaneously handle long-range text dependencies, breaking through the sequential processing bottleneck of recurrent neural networks (RNNs).59% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/687632API
curl https://kongchang.com/api/v1/knowledge/claims/687632MCP
get_claim(id=687632)