Search papers, labs, and topics across Lattice.
7
0
13
7
Vulnerabilities spanning multiple functions can be detected more accurately with VulAgentRL, which verifies evidence through a novel Code Property Graph approach.
Bio-SFT achieves superior HDR reconstruction by leveraging biological principles, outperforming conventional methods in both quality and efficiency.
Tibetan vision-language models can now leap forward thanks to FTibSuite, a new resource suite that closes the infrastructure gap and delivers a strong baseline with substantial performance gains.
Don't try to directly extrapolate wave fields to higher frequencies when data is scarce; instead, anchor on the transferable amplitude structure and separately recover the oscillatory details.
Agentic AI coding assistants are alarmingly susceptible to prompt injection attacks via external artifacts, effectively turning them into remote shells for attackers.
LLM agents are not just chatbots: they can find hundreds of real-world software vulnerabilities that traditional tools miss.
Security patch detectors trained on standard vulnerability databases are practically useless in the real world, losing up to 90% F1-score when deployed on in-the-wild data.
