Search papers, labs, and topics across Lattice.
6
0
5
8
Task-specific adaptation transforms LALMs from near-chance performers to competitive players in spoofing-aware speaker verification.
Evolved programs can exploit perceptual hash algorithms with unprecedented efficiency, exposing significant vulnerabilities in content moderation.
LLMs can crack 4x more passwords with automatically evolved prompts, revealing a significant vulnerability in password security.
Uncover the hidden vulnerabilities of your voice anti-spoofing model with a new tool that quantifies the probability of failure against unseen speech synthesis attacks.
Speech deepfake detection gets a reasoning upgrade: HIR-SDD uses chain-of-thought prompting with Large Audio Language Models to not only detect fakes but also explain *why* it thinks they're fake.
Sparse autoencoders, hyped as a key interpretability tool, may not be learning much more than random feature sets, casting doubt on their ability to decompose model internals.
