Chen Chen

CURE achieves up to 3.97% better performance than leading multimodal fusion methods while slashing computational costs by nearly 88%.

High variance in draft quality is tackled head-on, leading to a remarkable 66% increase in throughput for speculative decoding in large language models.

Optimizing conversational timing as a standalone objective can lead to more natural interactions without compromising reasoning abilities in dialogue systems.

TimeThink revolutionizes video reasoning by enabling models to pinpoint relevant temporal evidence with unprecedented accuracy, outperforming existing approaches.

DL-SLAM achieves a 13% boost in tracking accuracy by leveraging dual-level probabilistic frameworks to eliminate artifacts from dynamic objects in SLAM.

PULSE slashes communication overhead by 89% and boosts training throughput by up to 2.3x, revolutionizing how we scale diffusion models across GPU clusters.

Uncovering critical security and privacy vulnerabilities in foundation-model-powered robots could redefine how we approach their deployment in real-world applications.

MIXGUARD achieves robust privacy protection in split learning without sacrificing model utility, outperforming existing defenses against advanced data reconstruction attacks.

LLMs can exploit syntactic patterns to falsely inflate detection rates in hardware security benchmarks, but a new obfuscation framework can slash their effectiveness by up to 78.6%.

LLMs can now scale depth more effectively: a new attention mechanism recovers diluted features in deeper layers, boosting performance with negligible overhead.