Deep Learning
BackdoorAlign: Mitigating Fine-tuning based Jailbreak Attack with Backdoor Enhanced Safety Alignment
Alignment method inspired by an analogy with the concept of backdoor attacks. In particular, service providers will construct prefixed safety examples with a secret prompt, acting as a "backdoor trigger".
SLED: Self Logits Evolution Decoding for Improving Factuality in Large Language Models
From an optimization perspective, our SLED framework leverages the latent knowledge embedded within the LLM by contrasting the output logits from the final layer with those from early layers.
Feedback control guides credit assignment in recurrent neural networks
How do brain circuits learn to generate behaviour?
PV-Tuning: Beyond Straight-Through Estimation for Extreme LLM Compression
There has been significant interest in "extreme" compression of large language models (LLMs), i.e., to 1-2 bits per parameter, which allows such models to be executed efficiently on resource-constrained devices.
HairDiffusion: Vivid Multi-Colored Hair Editing via Latent Diffusion Y u Zeng
StyleGAN to address this task.
T ARSS-Net: T emporal-A ware Radar Semantic Segmentation Network
Radar signal interpretation plays a crucial role in remote detection and ranging.
Segment, Shuffle, and Stitch: A Simple Layer for Improving Time-Series Representations
However, non-adjacent sections of real-world time-series may have strong dependencies.
