Miyoshi Kasei (Japan)

Comprehensive gait characteristics on electromagnetic stance-control knee ankle foot orthosis in healthy individuals

Abstract Computer-controlled knee ankle foot orthosis (KAFO) has the potential to contribute to gait training and functional walking assistance by ensuring knee stability during stance while allowing knee flexion during swing; however, evidence regarding the gait characteristics...

Peri-Plate Gas Detected on Cervical CT as an Exploratory Imaging Indicator of Esophageal Perforation After Anterior Cervical Spine Surgery

STUDY DESIGN: Exploratory, single-center retrospective observational study. OBJECTIVE: To evaluate candidate imaging markers associated with esophageal perforation in patients undergoing reoperative anterior cervical exploration more than 1 month after anterior cervical spine sur...

Long-term effects of cognitive-behavioral therapy for adults with ADHD: results from an 8-year randomized controlled trial follow-up

Abstract Background To evaluate the provisional long-term durability of group cognitive behavioral therapy (CBT) benefits 8 years post-baseline and to explore tentative associations between digital booster engagement and the maintenance of treatment gains in adults with ADHD. Met...

Pull: Lazy Materialization of Working Memory for Stateful LLM Conversations

As LLM conversations grow to hundreds of turns, full-context injection incurs O(N²) cumulative token costs, while lossy summarization or hard truncation irreversibly discards historical state. We propose Pull, a session router that maintains an addressable metadata directory via...

Pull: Lazy Materialization of Working Memory for Stateful LLM Conversations

As LLM conversations grow to hundreds of turns, full-context injection incurs O(N²) cumulative token costs, while lossy summarization or hard truncation irreversibly discards historical state. We propose Pull, a session router that maintains an addressable metadata directory via...

KongMen: Real-Time Dialogue Purification via Training-Free Entity-Redundancy Gating

Long-context LLMs (128K–1M tokens) reduce truncation loss, but full-length inputs incur proportional latency, cost, and memory overhead. We propose KongMen, a training-free, LLM-free real-time dialogue purifier for multi-turn conversations. KongMen removes conversation-internal r...

KongMen: Real-Time Dialogue Purification via Training-Free Entity-Redundancy Gating

Long-context LLMs (128K–1M tokens) reduce truncation loss, but full-length inputs incur proportional latency, cost, and memory overhead. We propose KongMen, a training-free, LLM-free real-time dialogue purifier for multi-turn conversations. KongMen removes conversation-internal r...