כתבה
arXiv cs.LG ·
KuaFu: דחיסת התנהגות משתמש ארוכה
KuaFu: Compressing Long User Behavior into Understanding at Billion Scale
KuaFu הוא שכבת דחיסת התנהגות מאוחדת. היא מדחיסה כל פריט התנהגות ל-2-4 טוקנים. KuaFu עובדת על פלטפורמת הפרסום וההמלצות של Tencent.
תקציר מקורי באנגליתarXiv:2609.31045v1 Announce Type: cross Abstract: Conversational agents, generative recommenders, and personalized advertising all rest on one capability: understanding each user from raw behavior. Prevailing industrial practice is task-specific: for each task, a relevant subsequence is extracted from the full history and a dedicated model trained on it. In production it hits two bottlenecks. First, even after filtering, a single-task sequence stays extremely long: content-interest summarization reads several hundred items per user, tens of thousands of tokens once serialized as prompt text. Second, profiles are refreshed routinely: a billion users weekly, roughly 100K QPM in aggregate, which under a fixed GPU budget sets a hard throughput floor. Compression is therefore mandatory, yet tru
קרא במקור המקורי
arxiv.org
פתח כתבה מקורית