Full-Duplex-Bench
AI overview
Sign in and the AI will write an overview from our coverage.
Headlines · 2
- Orilo launches a voice-first personal AI agent
量子位 reports that Orilo is a voice-first personal agent that delivers personalized morning briefings, tracks long-tail information, and can handle email and calendar tasks through spoken commands. It uses a layered architecture with a low-latency speech model for interaction and a reasoning agent for background tool calls; in the cited Full-Duplex-Bench test, its reported interruption-recovery time was 1.08 seconds. The reviewer also noted that the initial product supports only English and sometimes reported completed actions when background integrations had failed.
量子位(原生 RSS) · 🔥 0 - HiPLEX: Hierarchical Policy Factorization for Full-Duplex Speech Language Models
The paper introduces HiPLEX, a reinforcement learning framework that factorizes a pretrained full-duplex text policy into a control policy deciding when to emit content (choosing among pad, epad and con) and a conditional content policy that picks a token only when con is selected. The authors report that across three Moshi seeds on Full-Duplex-Bench v1, HiPLEX reduces takeover rates during natural user pauses and backchannel opportunities and shortens post-interruption response latency versus GRPO, while keeping comparable judged interruption-response quality, and better matches pooled human turn-timing and backchannel-rate marginals on Moshi and PersonaPlex.
Hugging Face · Papers · 🔥 0
Experience and discussion from the community
Share my Full-Duplex-Bench experienceAsk about Full-Duplex-Bench
Nobody has shared their experience with Full-Duplex-Bench yet.