Dimensions overall score 8.0

GitHub Code Pulse

No public code linked for this paper yet.

• STEM-Bench is the first benchmark for streaming evaluation of memory in dialogue systems.technical 95%
• STEM-Bench comprises over 14K QA pairs in dialogue streams that assess perception fidelity, temporal reasoning, and global awareness.technical 90%
• Existing read-then-think memory mechanisms are fundamentally misaligned with infinite-horizon streaming dialogues.limitation 85%
• There exists a critical fidelity-efficiency dilemma where retrieval-based methods use fragment context while full-context models incur unbounded latency.limitation 90%
• ProStream enables ad-hoc memory recall on demand by reasoning over continuous streams with multi-granular distillation.method 85%
• ProStream employs Adaptive Spatiotemporal Optimization to dynamically optimize retention based on expected utility.method 90%
• ProStream enables bounded knowledge state for lower inference latency without sacrificing reasoning fidelity.result 85%
• Experiments show that ProStream outperforms baselines in both accuracy and efficiency.result 80%

Competitive landscape

Competitor map is still being generated for this paper. Enable generation or check back soon.

Startup potential card