BestHub
Sign in
Machine Learning Algorithms & Natural Language Processing
Machine Learning Algorithms & Natural Language Processing
Feb 9, 2026 · Artificial Intelligence

Time‑o1: Overcoming Time‑Series Forecasting Bottlenecks with a Novel Loss Function

The paper identifies two fundamental issues in time‑series forecasting—label autocorrelation bias and task‑scale explosion caused by the standard TMSE loss—and proposes Time‑o1, a PCA‑based orthogonal label transformation that eliminates bias, reduces optimization complexity, and yields consistent performance gains across multiple models and datasets.

NeurIPS 2025PCATime‑o1
0 likes · 12 min read
Time‑o1: Overcoming Time‑Series Forecasting Bottlenecks with a Novel Loss Function
Alibaba Cloud Big Data AI Platform
Alibaba Cloud Big Data AI Platform
Dec 23, 2025 · Artificial Intelligence

How Skrull Boosts Long-Context Fine‑Tuning Speed Up to 7.5×

The Skrull system, accepted at NeurIPS 2025, dynamically schedules long and short sequences during each training iteration, overlapping communication and computation to achieve up to 7.54× speedup for long‑context fine‑tuning of large language models while maintaining stability through load‑balancing and rollback mechanisms.

Dynamic Data SchedulingLong Context Fine-TuningModel Training Optimization
0 likes · 8 min read
How Skrull Boosts Long-Context Fine‑Tuning Speed Up to 7.5×
HyperAI Super Neural
HyperAI Super Neural
Dec 23, 2025 · Artificial Intelligence

NeurIPS 2025‑Selected Multi‑Stream Control Framework Achieves Precise Audio‑Visual Sync via Audio Demixing

The paper introduces a NeurIPS 2025‑selected multi‑stream video generation framework that demixes audio into speech, effects, and music, using dedicated control streams and a multi‑stage training strategy to achieve markedly better lip‑sync, event timing, and overall visual quality than prior methods.

MTV frameworkNeurIPS 2025audio demixing
0 likes · 9 min read
NeurIPS 2025‑Selected Multi‑Stream Control Framework Achieves Precise Audio‑Visual Sync via Audio Demixing
PaperAgent
PaperAgent
Nov 29, 2025 · Industry Insights

NeurIPS 2025 Insights: AI Agents, Reasoning, and the Shift to Real-World Systems

An analysis of the 5,984 papers accepted at NeurIPS 2025 shows a decisive move from ever‑larger models toward agents, reasoning‑focused LLMs, efficiency engineering, AI for Science, and trustworthy AI, signaling the transition from a research‑toy era to an engineering‑driven AI ecosystem.

AI for ScienceAI trendsLLM
0 likes · 7 min read
NeurIPS 2025 Insights: AI Agents, Reasoning, and the Shift to Real-World Systems