Unverified50% confidenceBenchmarkExact time
The short-term BTC/USDT price direction prediction system built by the developer uses a Transformer architecture combined with BiLSTM and cross-attention mechanisms, with about 1.5 million training samples, and the model's direction accuracy is about 60%, a clear improvement over the roughly 52% baseline.
1
Sources
50%
Confidence
Long-term
Relevance
10/12/2026
First Seen
Sources
Related Entities
Related Claims
Unverified2024年MIT等机构提出的测试时训练（TTT）架构尝试用线性注意力层替代部分Transformer层，使权重成为可在推理时快速更新的隐式记忆72% similarUnverifiedCurrently, over 80% of mainstream large models, including ChatGPT and DeepSeek, are trained based on Transformer or its variants.71% similarUnverifiedTransformer resolves the long-range dependency problem and improves training efficiency by computing attention weights between all positions in parallel, laying the foundation for large models such as GPT, BERT, and Claude.71% similarUnverifiedTransformer-based end-to-end tracking approaches have high requirements for training data and computing power, resulting in a steep deployment barrier.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1009425API
curl https://kongchang.com/api/v1/knowledge/claims/1009425MCP
get_claim(id=1009425)