AI Dynamics

Global AI News Aggregator

About

@jiqizhixin

  • Huawei’s Smart Noise Scheduler Improves Diffusion LLM Logic Learning
    Huawei’s Smart Noise Scheduler Improves Diffusion LLM Logic Learning

    By

    @jiqizhixin

    –

    What if Diffusion LLMs learned logic more efficiently? Huawei's Noah's Ark Lab proposes a breakthrough: their "smart noise scheduler" uses priority masking to focus training on only information-dense data, making DLLMs master core reasoning and structure. This boosts average accuracy by 4% on Code & Math reasoning, beating uniform baselines and unlocking DLLM potential. Mask Is What DLLM Needs: A Masked Data Training Paradigm for Diffusion LLMs Paper: arxiv.org/abs/2603.15803 Dataset: huggingface.co/datasets/malr… Our report: mp.weixin.qq.com/s/1yTd36hev… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-05 14:05 UTC

  • HiFi-Inpaint: ByteDance AI Framework for Detail-Preserving Product Images
    HiFi-Inpaint: ByteDance AI Framework for Detail-Preserving Product Images

    By

    @jiqizhixin

    –

    How do you get AI to create stunning product images without losing crucial details? ByteDance and a collaboration of top universities present HiFi-Inpaint. This novel AI framework employs 'Shared Enhancement Attention' to meticulously refine fine-grained product features and 'Detail-Aware Loss' for pixel-perfect guidance, supported by a new large-scale dataset, HP-Image-40K. HiFi-Inpaint achieves state-of-the-art performance, generating human-product images with unprecedented detail preservation, set to transform digital marketing and e-commerce visuals. HiFi-Inpaint: Towards High-Fidelity Reference-Based Inpainting for Generating Detail-Preserving Human-Product Images Paper: arxiv.org/abs/2603.02210  Project: correr-zhou.github.io/HiFi-I… Our report: mp.weixin.qq.com/s/xoIJU4fBc… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-05 07:02 UTC

  • SWE-MiniSandbox: Container-Free RL for Software Engineering Agents
    SWE-MiniSandbox: Container-Free RL for Software Engineering Agents

    By

    @jiqizhixin

    –

    What if you could train AI software engineers faster, without the heavy overhead of containers? Researchers from Peking University, Ant Groupe, and The University of Hong Kong present SWE-MiniSandbox. This novel, container-free method uses kernel-level isolation and lightweight pre-caching, eliminating bulky container images for reinforcement learning. It achieves comparable performance to container-based pipelines while reducing disk usage by 95% and environment setup time by 75%, making scalable RL training far more accessible for software engineering agents. SWE-MiniSandbox: Container-Free Reinforcement Learning for Building Software Engineering Agents Paper: arxiv.org/abs/2602.11210 Code: github.com/lblankl/SWE-MiniS… Docs: lblankl.github.io/SWE-MiniSa… Our report: mp.weixin.qq.com/s/NlQLprZmM… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-05 04:00 UTC

  • STAGE: AI Multi-shot Video Generation with Narrative Consistency
    STAGE: AI Multi-shot Video Generation with Narrative Consistency

    By

    @jiqizhixin

    –

    Can AI finally create multi-shot videos with seamless narrative flow and cinematic flair? Researchers from Beijing University of Posts and Telecommunications and Peking University present STAGE. This new method rethinks video generation by planning full shot-by-shot storyboards using start-to-end frame pairs. It employs smart memory and clever encoding to keep characters and scenes consistent, ensuring smooth visuals within and between shots. STAGE significantly outperforms existing methods, achieving superior narrative control and unparalleled visual consistency across cinematic sequences. STAGE: Storyboard-Anchored Generation for Cinematic Multi-shot Narrative Paper: arxiv.org/abs/2512.12372 Code: github.com/escapistmost/Stor… Our report: mp.weixin.qq.com/s/rmeF2tbIu… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-04 18:55 UTC

  • EasySteer: Unified Framework for High-Performance LLM Steering
    EasySteer: Unified Framework for High-Performance LLM Steering

    By

    @jiqizhixin

    –

    Want to precisely control your LLM's behavior without expensive retraining? New research from Zhejiang University unveils EasySteer, a unified framework for high-performance and extensible LLM steering. It's a unified framework that lets you finely tune LLM responses in real-time by subtly adjusting their internal 'thoughts' or hidden states. This lightweight method offers modular control and pre-computed steering options, sidestepping costly model retraining. EasySteer achieves a game-changing 10.8-22.3x speedup over current methods. It dramatically reduces common LLM issues like overthinking and hallucinations, making advanced steering a robust, production-ready tool for deployable, controllable language models. EasySteer: A Unified Framework for High-Performance and Extensible LLM Steering Paper: arxiv.org/abs/2509.25175 Code: github.com/ZJU-REAL/EasyStee… Our report: mp.weixin.qq.com/s/dxuJHvXOf… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-04 14:53 UTC

  • Error-Entropy Scaling Law Surpasses Traditional Cross-Entropy for LLM Development
    Error-Entropy Scaling Law Surpasses Traditional Cross-Entropy for LLM Development

    By

    @jiqizhixin

    –

    Is the fundamental scaling law guiding large language model development broken? Researchers from Tsinghua University have found the answer. They've decomposed cross-entropy loss into three components: Error-Entropy, Self-Alignment, and Confidence, finding that only Error-Entropy truly scales with model size. This new "Error-Entropy scaling law" provides a far more accurate guide for LLM development, outperforming the traditional cross-entropy law, especially for the largest models. Crucial for future AI design. What Scales in Cross-Entropy Scaling Law? Paper: arxiv.org/abs/2510.04067 Code: github.com/yanjx2021/Rethink… Our report: mp.weixin.qq.com/s/ngn6YY6Aj… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-04 08:49 UTC

  • Shop-R1: AI Framework for Understanding Human Online Shopping Behavior
    Shop-R1: AI Framework for Understanding Human Online Shopping Behavior

    By

    @jiqizhixin

    –

    Ever wonder if an AI could truly understand how you shop online? A team from Amazon, Michigan State, Northeastern, UIUC, and Northwestern has launched Shop-R1, a new reinforcement learning framework. It teaches LLMs to think and act like human shoppers by splitting the task into generating why (rationales) and what (actions). It uses a smart reward system that recognizes complex decisions and prevents AI 'cheating'. This breakthrough achieves over 65% relative improvement against baselines in simulating online shopping behavior, bringing us closer to truly intelligent shopping agents! Shop-R1: Rewarding LLMs to Simulate Human Behavior in Online Shopping via Reinforcement Learning Paper: arxiv.org/abs/2507.17842 Project: damon-demon.github.io/shop-r… Our report: mp.weixin.qq.com/s/Dvst0Oirm… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-04 05:43 UTC

  • Qwen 3.6 Plus Breaks 1 Trillion Tokens Record on OpenRouter
    Qwen 3.6 Plus Breaks 1 Trillion Tokens Record on OpenRouter

    By

    @jiqizhixin

    –

    Huge accomplishment! Congrates to @Alibaba_Qwen team. OpenRouter (@OpenRouter) Qwen 3.6 Plus from @Alibaba_Qwen is officially the first model on OpenRouter to break 1 Trillion tokens processed in a single day! At ~1,400,000,000,000 tokens, it’s the strongest full day performance of any new model dropped this year. Congrats to the Qwen team! — https://nitter.net/OpenRouter/status/2040239467865489874#m

    → View original post on X — @jiqizhixin, 2026-04-04 02:47 UTC

  • Intelligent Remote Sensing Agents Transform Earth Observation with AI
    Intelligent Remote Sensing Agents Transform Earth Observation with AI

    By

    @jiqizhixin

    –

    What if Earth observation could truly think for itself? A collaborative team from Hong Kong University of Science and Technology, Northwestern Polytechnical University, Tsinghua University, and international partners have released a seminal survey on "Intelligent Remote Sensing Agents." This new paradigm shows how AI agents integrate perception, planning, memory, and tool execution to autonomously achieve complex geospatial understanding. This transforms remote sensing from passive data collection into proactive, intelligent decision support, far surpassing previous capabilities in urban governance, precision agriculture, ecological monitoring, and emergency response. Intelligent Remote Sensing Agents: A Survey Paper: github.com/PolyX-Research/Aw… Repo: github.com/PolyX-Research/Aw… Our report: mp.weixin.qq.com/s/QYAyTjaAa… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-04 01:40 UTC

  • MoGraphGPT: Creating Interactive 2D Scenes Without Coding

    By

    @jiqizhixin

    –

    What if building complex, interactive 2D scenes was as simple as describing them, with no coding needed? Enter MoGraphGPT! This new system leverages modular LLMs, using specialized AIs for individual scene elements and a central AI to manage their interactions, all through an intuitive graphical user interface with auto-generated sliders. It offers precise visual control for scene creation. MoGraphGPT significantly outperforms Cursor Composer, making the creation of multi-element 2D interactive scenes easier, more controllable, and higher performing, all without writing a single line of code. MoGraphGPT: Creating Interactive Scenes Using Modular LLM and Graphical Control Paper: ieeexplore.ieee.org/abstract… Our report: mp.weixin.qq.com/s/objKgAzNO… 📬 #PapersAccepted by Jiqizhixin

    → View original post on X — @jiqizhixin, 2026-04-03 19:50 UTC