Oceania
On Reinforcement Learningand Distribution Matchingfor Fine-Tuning Language Models withno Catastrophic Forgetting
Twoofthemcanbecharacterizedas "Reward Maximization" (RM): Standard Policy Gradients (PG) and KL-control.
Inherently Explainable Reinforcement Learning in Natural Language
Observation: Up a tree Beside you on the branch is a small birds nest In the birds nest is a large egg encrusted with precious jewels, scavenged by a childless songbird... Explanation: I am in the Forest Path now.
Not All Out-of-Distribution Data Are Harmful to Open-Set Active Learning Y ang
Recall (higher is better): the ratio of selected ID instances to the total number of ID instances.
Not All Out-of-Distribution Data Are Harmful to Open-Set Active Learning Y ang
Recall (higher is better): the ratio of selected ID instances to the total number of ID instances.
Improvedtechniquesfordeterministicl2robustness
Gradient NormPreserving (GNP) architectures where each layer preserves the gradient norm during backpropagation. For 1-Lipschitz Convolutional Neural Networks (CNNs), this involves using orthogonal convolutions (convolution layers with an orthogonal Jacobian matrix) [Li et al., 2019b, Trockman and Kolter,
