Unverified50% confidenceFactExact time
Progressive Neural Networks add a new column of network layers for each new task, routing features from old columns into the new column via lateral connections while freezing the old columns' parameters, so their parameter count grows linearly with the number of tasks.
1
Sources
50%
Confidence
Long-term
Relevance
10/27/2026
First Seen
Sources
Related Entities
Related Claims
Unverified渐进式神经网络为每个新任务分配独立的网络列,代价是模型参数量随任务数线性增长80% similarUnverified卷积神经网络的核心创新在于权重共享与局部连接机制,通过滑动卷积核共享同一组参数降低模型参数量70% similarUnverifiedParameters in large models are mathematically the weight values and bias values on each connection edge of a neural network, randomly initialized before training and continuously adjusted through backpropagation and gradient descent.70% similarUnverifiedIn the class-incremental graph learning scenario of graph neural networks, as new classes arrive one after another the model exhibits catastrophic forgetting, and predictions are often dominated by the most recently learned classes, producing recency bias.70% similarUnverifiedNeural cellular automata replace the update rules of traditional cellular automata with trainable neural networks, enabling the evolution of cell states to be learned from data.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1102234API
curl https://kongchang.com/api/v1/knowledge/claims/1102234MCP
get_claim(id=1102234)