Unverified50% confidenceBenchmarkExact time
By 2019, researchers were able to break reCAPTCHA v2 image-selection challenges with 70% accuracy using convolutional neural networks.
1
Sources
50%
Confidence
Long-term
Relevance
9/9/2026
First Seen
Sources
Related Entities
Related Claims
Unverified该C++手写CNN项目包含完整的卷积神经网络结构、反向传播、梯度检查和随机梯度下降等核心模块，最终在验证集上达到约94%的准确率68% similarUnverified鼠标动力学认证技术仅凭鼠标移动曲线即可识别用户身份，准确率在部分研究中超过95%64% similarUnverifiedThe tested task was conducted over four rounds, dispatching a cumulative total of 68 sub-agents to complete batch prompt testing for a digital human writing skills module.61% similarUnverifiedA GraphRAG-related paper published by Microsoft Research in 2024 validated that graph-augmented methods significantly outperform pure vector retrieval when handling complex questions requiring cross-document reasoning.61% similarUnverifiedThe experiment targeted a brushless DC (BLDC) actuator, with 120 agents distributed across 4 automated research sessions.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/883930API
curl https://kongchang.com/api/v1/knowledge/claims/883930MCP
get_claim(id=883930)