Accurate CPU-GPU Latency Estimation for Mobile Edge DVFS
Discover FLAME, a tool for precise CPU-GPU latency estimation in mobile edge devices, optimizing DVFS and reducing profiling time drastically.
Sequential KV Cache Compression Beyond Shannon Limit
Discover advanced sequential KV cache compression using probabilistic tries, surpassing the per-vector Shannon limit for transformer models.
SocialWise: AI Therapy to Boost Autism Communication Skills
SocialWise uses AI-driven conversation therapy to help individuals with autism improve communication skills through interactive, real-time feedback.
LLMs in Design & Dev: Tools or Teammates?
Explore how designers and developers navigate using LLMs as tools or teammates, balancing roles, accountability, and collaboration in workflows.
Architectural Limits and Design Challenges in Human-LLM Systems
Explore the architectural limits, metacognitive co-option, and design challenges in human-LLM systems to enhance user autonomy and cognitive integrity.