Notable data and AI papers

Papers627Labs45

627 papers · 45 labs

On-Policy Self-Training: Data-Free Self-Improvement in Large Language Models research paper by Google, 2026

Google · Training and scaling · Oct 23, 2026 · 31 upvotes · unverified

Read on arXiv

What it shows

Foundation models can self-improve across text and vision tasks by generating synthetic questions and reasoning traces guided only by high-level topics, without external supervision.

By Leon Sick, Lukas Hoyer, Yannick Strümpler and 3 more · arXiv 2610.32770 · PDF

UnverifiedHugging Face's summary; not yet checked by hand.

More from Google

All 45
Inference and efficiencyUnverifiedOct 27, 2026EmbeddingGemma 2: Unified and Efficient Multimodal RepresentationsEmbeddingGemma 2 is an open multimodal embedding model that unifies text, images, audio, and video representations with efficient resource-flexible design and state-of-the-art benchmark performance.Google34 upvotesRead →Multimodal and roboticsUnverifiedOct 26, 2026StreamVibe: Proactive Video Agents for Evolving User IntentStreamVibe introduces a benchmark, streaming vision-language model, and hierarchical memory harness to enable proactive, intent-aware interaction with implicit and evolving user goals.Google78 upvotesRead →Training and scalingUnverifiedOct 14, 2026DeLoop: Repeated Actions as Self-Supervision for Long-Horizon AgentsThis self-supervised training method teaches AI agents to avoid repetitive, unproductive actions during long tasks by using their own history as feedback.Google21 upvotesRead →Agents and evaluationUnverifiedOct 13, 2026CLASHarness: Benchmarking Multi-Agent Conflict in Computer Use via a Controllable and Verifiable Generation PipelineThis paper introduces a benchmark and generation pipeline to evaluate how agents resolve conflicts and verify information in multi-agent computer-use scenarios.Google36 upvotesRead →Agents and evaluationUnverifiedOct 9, 2026Harness^2: Recursive Agent Harnessing for an Open WorldThis paper introduces a recursive framework that allows agents to automatically improve their own execution harnesses for better performance on novel open-world tasks.Google2 citations123 upvotesRead →Multimodal and roboticsUnverifiedOct 7, 2026VIS-Ground: Video Interactive Storytelling with Contextual GroundingThis paper introduces a framework that transforms heterogeneous user inputs into structured constraints to improve grounding in interactive video generation.Google103 upvotesRead →Applied AIUnverifiedOct 2, 2026SEER: Self-Evolving Event Reasoning and Retrieval for Time Series ForecastingSEER improves time series forecasting by using a closed-loop system to retrieve and reason about external events that influence data trends.Google35 upvotesRead →Retrieval and dataUnverifiedOct 1, 2026RPTune: Learned Context Curation for LLM Catalog SearchRPTune improves LLM search accuracy by learning how to curate product catalogs and adapting the model to those curated contexts.Google40 upvotesRead →

Other training and scaling papers

Topic
Training and scalingJun 17, 2021LoRA: Low-Rank Adaptation of Large Language ModelsLoRA fine-tunes a large model by training small low-rank matrices, cutting trainable parameters by 10,000 times.Microsoft25.6k citations66 upvotesRead →Training and scalingJan 23, 2020Scaling Laws for Neural Language ModelsLanguage model loss falls as a smooth power law as model size, data and compute grow.OpenAI9,592 citations12 upvotesRead →Training and scalingMar 29, 2022Training Compute-Optimal Large Language ModelsChinchilla: for a fixed compute budget, train a smaller model on more data; parameters and tokens should grow together.Google DeepMind4,035 citations13 upvotesRead →Training and scalingSep 17, 2019Megatron-LM: Training Multi-Billion Parameter Language Models Using Model ParallelismMegatron-LM splits each Transformer layer across GPUs to train models with billions of parameters.NVIDIA3,387 citations5 upvotesRead →Training and scalingUnverifiedJan 5, 2024DeepSeek LLM: Scaling Open-Source Language Models with LongtermismDeepSeek LLM, an open-source language model project, develops a large dataset and employs SFT and DPO to achieve performance surpassing LLaMA-2 70B and GPT-3.5 in various benchmarks and open-ended evaluations.DeepSeek872 citations56 upvotesRead →Training and scalingMay 15, 2024LoRA Learns Less and Forgets LessLoRA learns less than full fine-tuning on code and math, but forgets less of what the model already knew.Databricks486 citations91 upvotesRead →
About this paper
Citations
Not counted yet · Semantic Scholar
Upvotes
31 · Hugging Face
Lab
Google · on Companies · on Acquisitions · on Paydays · on TechConf · on Releases

Changes

UpvotesNov 9, 2026Hugging Face upvotes fell from 48 to 31.
New paperNov 2, 2026Found by the weekly Hugging Face scan; not yet checked by hand.