الملخص
While Large Language Models (LLMs) advertise million-token context windows, reasoning quality often collapses as inputs grow -- a phenomenon termed context rot. This failure stems from a structural entanglement in monolithic architectures, where the massive search burden of contextual grounding exhausts the representational capacity needed for complex reasoning. To resolve this, we propose Grounding-Reasoning Disaggregation via DIStributed long COntext scaling (DISCO). Inspired by distributed computing frameworks like Apache Spark, DISCO partitions long context across a fleet of Worker LLMs dedicated exclusively to parallel, localized grounding. A central Driver LLM, trained via Reinforcement Learning (GRPO) to optimize planning, orchestrates execution by dynamically mapping queries into atomic extraction tasks and reducing the gathered evidence to synthesize a final answer. By isolating reasoning from raw context noise, DISCO effectively eliminates context rot. On RULER-QA (1M tokens), it maintains 78.4% accuracy where standard baselines collapse. Furthermore, it outperforms full-context models by up to 9.8 points on LongBench v2 and matches frontier models like Gemini-3-Pro-Preview while reducing inference costs by over 80%, establishing a highly efficient paradigm for robust long-context inference.
الكلمات المفتاحية
الموضوع
بيانات النشر
- المجلة
- غير متاح
- وصول مفتوح
- وصول مفتوح أخضر
اقتبس هذه المقالة
APA 7
Chen, G., Lai, V. D., Mukherjee, S., Kveton, B., Yoon, S., Dernoncourt, F., Xie, Q., & Bui, T. (2026). DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation. https://omanscience.com/ar/articles/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation
MLA 9
Chen, Guanzheng, et al. "DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation." https://omanscience.com/ar/articles/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.
شيكاغو (المؤلف–التاريخ)
Chen, Guanzheng, Viet Dac Lai, Subhojyoti Mukherjee, Branislav Kveton, Seunghyun Yoon, Franck Dernoncourt, Qizhe Xie, and Trung Bui. 2026. "DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation." https://omanscience.com/ar/articles/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.
هارفارد
Chen, G., Lai, V. D., Mukherjee, S., Kveton, B., Yoon, S., Dernoncourt, F., Xie, Q. and Bui, T. (2026) 'DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation', Available at: https://omanscience.com/ar/articles/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.
فانكوفر
Chen G, Lai VD, Mukherjee S, Kveton B, Yoon S, Dernoncourt F, et al. DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation. https://omanscience.com/ar/articles/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation
IEEE
G. Chen, V. D. Lai, S. Mukherjee, B. Kveton, S. Yoon, F. Dernoncourt, Q. Xie, and T. Bui, "DISCO: Distributed Long Context Scaling with Grounding-Reasoning Disaggregation," https://omanscience.com/ar/articles/disco-distributed-long-context-scaling-with-grounding-reasoning-disaggregation.