Tuesday, August 18, 2026
Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination; Self-Supervised Visual On-Policy Distillation; Beyond Final Scores
Executive Summary
Today's HuggingFace trending papers (from 2026-08-18) are led by Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination (259 upvotes), which existing benchmarks, however, provide limited evidence on detector and generator behavior in such settings, including how detectability varies with ge. Close behind is Self-Supervised Visual On-Policy Distillation (148 upvotes).
Key themes today: Autonomous Agents, Reasoning & Chain-of-Thought, Model Architecture.
On the model side, Qwen3.8-27B by Qwen leads trending models with 415,039 downloads.
GitHub trending highlights: public-apis/public-apis (1907 stars today), harry0703/MoneyPrinterTurbo (1189 stars today), cordiverse/cordis (957 stars today).
Researcher Notes
Top paper: Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination. Tagged [benchmarks, video, reinforcement-learning] with 259 upvotes. Recent video generators can fabricate realistic depictions of wars, disasters, public emergencies, and other real-world crises, creating substantial risks of misinformation. Existing benchmarks, howev
High-relevance papers: Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination; Self-Supervised Visual On-Policy Distillation. These had the highest community engagement and likely represent the day's most impactful contributions.
Rising themes: Autonomous Agents, Reasoning & Chain-of-Thought, Model Architecture. Multiple papers cluster around these topics, suggesting active research momentum.
Model leaderboard dominated by: Qwen (3 models), MiniMaxAI (2 models), unsloth (1 models).
GitHub spotlight: public-apis/public-apis (1907 stars today) — A collective list of free APIs
Themes & Trends
Autonomous Agents
risingSystems that plan, act, and improve autonomously — spanning coding agents, tool-use orchestration, and self-evolving architectures.
Reasoning & Chain-of-Thought
risingAdvances in structured reasoning, chain-of-thought prompting, and logical inference capabilities of language models.
Model Architecture
risingNovel architectures, scaling strategies, and training paradigms for foundation models.
Knowledge Distillation
risingMethods for transferring capabilities between models — including self-distillation, on-policy approaches, and compression techniques.
Evaluation & Benchmarks
risingNew benchmarks, evaluation methodologies, and analysis of existing evaluation frameworks.
Trending Papers (13)
Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination
High RelevanceShuo Liang, Yixing Ma, Pengfei Zhou, Xingyan Chen, Zihan Mei, Manting Li — Independent Research
Recent video generators can fabricate realistic depictions of wars, disasters, public emergencies, and other real-world crises, creating substantial risks of misinformation. Existing benchmarks, however, provide limited evidence on detector and generator behavior in such settings, including how detectability varies with generation conditions, how people perceive generated videos, and whether de...
Key Findings
- •
Existing benchmarks, however, provide limited evidence on detector and generator behavior in such settings, including how detectability varies with ge
- •
To address this gap, we introduce RA-Bench, a benchmark for AI-generated video detection that uses Real videos as Anchors.
- •
RA-Bench contains 17,886 videos, comprising 1,830 real-video anchors across 10 social-risk categories and 16,056 generated clips from four open-source
Self-Supervised Visual On-Policy Distillation
High RelevanceYijiang Li, Yijun Liang, Yunjie Tian, Bingyang Wang, Ke Zhang, Zhenfei Yin — Independent Research
Visual on-policy distillation relies heavily on an informative teacher-student asymmetry, through either a larger, stronger teacher or privileged supervision, such as reference answers or ground-truth regions of interest. This raises a fundamental question: where can informative asymmetry come from when nothing privileged is available? We answer this by inverting where the asymmetry comes from.
Key Findings
- •
This raises a fundamental question: where can informative asymmetry come from when nothing privileged is available?
- •
We answer this by inverting where the asymmetry comes from.
- •
Rather than adding privileged information to the teacher, we subtract information from the student.
Beyond Final Scores: A Systematic Evaluation of Agents for Long-Horizon AI Research and Development
Yiwei Li, Wanli Yang, Hexiang Tan, Xiangzhou Huang, Zhengyu Chen, Ziran Li — Independent Research
Autonomous agents are increasingly capable of improving models, systems, and other technical artifacts through long-horizon experimentation. To understand the current state of this capability, however, evaluation must go beyond final scores, which neither reveal where progress is gained or lost nor indicate whether accumulated experience improves later decisions. We therefore present a systemat...
Key Findings
- •
To understand the current state of this capability, however, evaluation must go beyond final scores, which neither reveal where progress is gained or
- •
We therefore present a systematic evaluation of seven frontier models on 36 long-horizon tasks based on a new framework that uses rule-based metrics t
- •
The results show that current agents operate more like engineering optimizers than fully autonomous researchers: they can formulate and implement prac
Intern-S2-Mobius: Foundation Model with Decoupled Knowledge and Reasoning
Kai Chen, Jifeng Ding, Ning Ding, Jiaye Ge, Lixin Gu, Yicheng Gu — Independent Research
We introduce Mobius-v0, an architecture that comprises a globally shared Memory (FFN) that stores knowledge vectors and multiple Reasoners (Self-Attn) that iteratively achieve compositional reasoning. Using hidden states as cache and carrier, reasoners repeatedly query memory for required knowledge-vectors, while the knowledge is transmitted back to reasoning operators. Through this knowledge-r...
Key Findings
- •
Using hidden states as cache and carrier, reasoners repeatedly query memory for required knowledge-vectors, while the knowledge is transmitted back to
- •
Through this knowledge-reasoning-separation architecture, Mobius achieves better knowledge compression and reasoning efficiency.
- •
Built upon Mobius-v0 architecture: 1) Our 7B model trained-from-scratch achieves similar downstream score as a 7B Transformer baseline with 62.6% of b
Apodex Discovery: Reality Benchmarks and Environments for Evaluating and Building Discoverative Artificial Intelligence
Brian Wang, Bin Feng, Xiaoman Pan, Chenyang An, Felix Liu, Tangqi Fang — Independent Research
Apollo did not reach the Moon merely because its engineers could solve difficult equations. It succeeded by turning a distant ambition into a mission architecture of explicit objectives, simulation, verification, and repeated correction. AI now faces a similar transition: frontier models can solve difficult tasks once the problem, tools, and success criteria are specified, yet consequential rea...
Key Findings
- •
It succeeded by turning a distant ambition into a mission architecture of explicit objectives, simulation, verification, and repeated correction.
- •
AI now faces a similar transition: frontier models can solve difficult tasks once the problem, tools, and success criteria are specified, yet conseque
- •
We introduce Apodex Discovery, a framework for building and evaluating discoverative AI through the heavy-duty solver, a system comprising a foundatio
SimpleOPD: Simple Tokenizer-Agnostic On-Policy Distillation for Long-Context Reasoning
Haonan He, Haodi Lei, Yun Luo, Haoran Zhang, Shunkai Zhang, Yizhuo Li — Independent Research
On-policy distillation (OPD) offers a promising way to transfer reasoning capabilities from stronger teacher models, but applying it to long-context reasoning teachers and short-context students introduces practical challenges, including tokenizer mismatch, teacher-student distribution mismatch, response length explosion, and training instability. In this work, we study this setting by transfer...
Key Findings
- •
In this work, we study this setting by transferring proof-reasoning capabilities from the long-context reasoning model SU-01 to short-context student
- •
To handle tokenizer differences, we perform OPD in a shared text space and align only tokens that occupy identical text spans under the student and te
- •
To mitigate the problem of excessive generation length and frequent truncation, we introduce a student reference KL loss and mask the advantages of sp
Marionette: Predicting World States, Rendering Geometry, Painting Appearance
Zian Meng, Zhen Li, Chuanhao Li, Qiang Li, Kaipeng Zhang — Independent Research
Interactive game world models typically autoregress visual observations directly in pixel or latent space, forcing structured properties such as pose, geometry, and occlusion to be implicitly maintained by the same generative sequence. Over long horizons, errors in these latent world properties accumulate, making consistency and controllability fragile. We explicitly model the evolving world st...
Key Findings
- •
Over long horizons, errors in these latent world properties accumulate, making consistency and controllability fragile.
- •
We explicitly model the evolving world state, delegate exact geometric computation to a fixed, zero-parameter renderer, and leave the neural model to
- •
We instantiate this idea as Marionette, a world model for interactive games with articulated characters.
DFM Mimir v1: An Open HRM Delivering Frontier Performance at 1B Parameters Using Only Permissible Post-Training Data
Peter Schneider-Kamp, Jacob Nielsen, Gianluca Barmina, Kenneth Enevoldsen, Lukas Galke Poech — Independent Research
Current large language model development relies on massive, often non-permissible datasets, creating a high barrier for researchers committed to open-source and ethically sourced data. We introduce Mimir v1, a 1-billion-parameter language model based on the Hierarchical Reasoning Model (HRM) architecture, that is trained from scratch and delivers highly competitive performance for English and s...
Key Findings
- •
We introduce Mimir v1, a 1-billion-parameter language model based on the Hierarchical Reasoning Model (HRM) architecture, that is trained from scratch
- •
Trained on a mixture of 161 datasets, Mimir v1 outperforms the original HRM-Text 1B and competes with larger frontier models like Qwen 3.5 4B and Gemm
- •
The model is available on the Hugging Face Hub: https://huggingface.co/danish-foundation-models/DFM-Mimir
MobileMem: Learning from a Year of Mobile Experiences
Xinle Deng, Yida Xue, Xiangyuan Ru, Haoming Xu, Shuofei Qiao, Mengru Wang — Independent Research
The next generation of AI agents is increasingly moving beyond systems that answer isolated questions toward persistent personal assistants that can understand, remember, and continuously learn from users' experiences. Such assistants require long-term memory to accumulate and leverage user-specific experiences over time, yet existing benchmarks remain inadequate for realistic mobile settings, ...
Key Findings
- •
Such assistants require long-term memory to accumulate and leverage user-specific experiences over time, yet existing benchmarks remain inadequate for
- •
We introduce MobileMem, a benchmark and framework for studying on-device long-term memory, grounded in a year-scale collection of mobile experiences.
- •
MobileMem employs a knowledge-grounded synthesis pipeline to construct coherent and temporally consistent long-horizon trajectories from user-app sess
HumanTracker: Towards Comprehensive and Human-Aligned Motion Tracking Benchmark
Dairu Liu, Zekun Qi, Jiayu Zeng, Ruixi Yu, Yu Guan, Yintianrun Zhang — Independent Research
Humanoid motion tracking is central to teleoperation and whole-body imitation, yet evaluation often disagrees with what people perceive in videos. Kinematic errors average per-frame pose differences but miss the physical artifacts that matter most, particularly unstable support and incorrect contacts such as foot skating and mistimed touch-downs. Meanwhile, widely used test suites are small and...
Key Findings
- •
Kinematic errors average per-frame pose differences but miss the physical artifacts that matter most, particularly unstable support and incorrect cont
- •
Meanwhile, widely used test suites are small and lack the diversity needed to stress contact-rich, long-horizon behaviors.
- •
We introduce HumanTracker to make humanoid tracking evaluation both perceptually aligned and scalable.
Latent On-Policy Self-Distillation
Guibin Zhang, Jiayang Lyu, Ran Sun, Xinlei Yu, Haoyu Zhao, Qibing Ren — Independent Research
Enabling agents to learn from experience and internalize it into their policy has become a central problem in self-evolving AI. On-policy self-distillation (OPSD) offers an effective pathway by using a privileged self-teacher to provide dense supervision on the student's own trajectories; however, existing methods still rely heavily on designer-specified privileged artifacts (e.g., answers, fee...
Key Findings
- •
On-policy self-distillation (OPSD) offers an effective pathway by using a privileged self-teacher to provide dense supervision on the student's own tr
- •
In this work, we introduce Latent On-Policy Self-Distillation (LOPD), which, rather than proposing another hand-crafted OPSD variant with a newly pres
- •
Technically, LOPD retrieves relevant experiences and composes them into continuous latent tokens that condition a self-teacher, while the student gene
CPI-Bench: A Comprehensive,Practical and Intelligent Benchmark for Real-World Image Editing
Qinye Zhou, Jun Zheng, Yongchao Du, Yuan Wang, Zhengrui Chen, Zuan Gao — Independent Research
With the rapid advancement of image editing models and their widespread application across various domains, there is an increasingly urgent need to deploy these model capabilities directly into real-world scenarios. However, existing benchmarks remain confined to simple single-image tasks, suffering from limited coverage dimensions and an inability to effectively differentiate performance among...
Key Findings
- •
However, existing benchmarks remain confined to simple single-image tasks, suffering from limited coverage dimensions and an inability to effectively
- •
Consequently, they fail to reliably evaluate model performance in complex multi-image editing, highly demanding reasoning instructions, and practical
- •
To address these limitations, we propose CPI-Bench, a Comprehensive, Practical andIntelligent benchmark for real-world image editing.
Modular Cognitive Architecture Emerges in Large Language Models
Pengrui Han, Jacob Andreas, Evelina Fedorenko, Andrea Gregor de Varda — Independent Research
The human brain exhibits a striking degree of functional specialization, with distinct networks supporting language, formal reasoning, reasoning about other minds, and reasoning about the physical world. Is this modular organization a fundamental principle of how intelligent systems must be built, or an evolutionary accident specific to biological brains? Here, we test whether a similar organiz...
Key Findings
- •
Is this modular organization a fundamental principle of how intelligent systems must be built, or an evolutionary accident specific to biological brai
- •
Here, we test whether a similar organization emerges in Large Language Models--another class of intelligent systems created through a very different o
- •
Using circuit analyses across N=46 tasks spanning four cognitive domains (language, formal reasoning, social reasoning, physical reasoning), we find t
Trending Models (10)
Qwen · image-text-to-text · Unknown
Qwen3.8-27B by Qwen. 415,039 downloads, 10,732 likes on HuggingFace.
unsloth · text-generation · Unknown
Qwen3.8-27B-GGUF by unsloth. 2,727,609 downloads, 1,639 likes on HuggingFace.
Qwen · text-generation · Unknown
Qwen3.8-2.4T-A95B by Qwen. 9,465 downloads, 1,042 likes on HuggingFace.
Lightricks · video-generation · Unknown
LTX-2.5 by Lightricks. 465,529 downloads, 1,114 likes on HuggingFace.
MiniMaxAI · audio-generation · Unknown
MiniMax-Music3 by MiniMaxAI. 10,375 downloads, 907 likes on HuggingFace.
meta-models · image-text-to-text · Unknown
Muse-Glimmer-30B by meta-models. 334,099 downloads, 1,662 likes on HuggingFace.
MiniMaxAI · video-generation · Unknown
MiniMax-H3 by MiniMaxAI. 2,403,238 downloads, 4,088 likes on HuggingFace.
deepseek-ai · text-generation · Unknown
DeepSeek-V4-Pro-0813 by deepseek-ai. 25,006 downloads, 574 likes on HuggingFace.
Qwen · image-text-to-text · Unknown
Qwen3.8-27B-FP8 by Qwen. 495,646 downloads, 532 likes on HuggingFace.
orcarouter · image-text-to-text · Unknown
Qwen3.8-27B-Uncensored-FP8 by orcarouter. 15,812 downloads, 444 likes on HuggingFace.
Trending GitHub Repos (12)
利用 AI 大模型和自动化工作流，根据主题或关键词一键生成高清短视频。Generate HD short videos from a topic or keyword with an automated AI workflow.
Meta-Framework of Spatiotemporal Composability
Local UI to run and train LLMs and diffusion models, including Qwen3.8, Kimi K3, MiniMax-H3, Gemma 4, DeepSeek-V4, FLUX and more.
14MB foundation model for tiny devices; phones, wearables, smart home, and robots.
Open-source AI penetration testing tool to find and fix your app’s vulnerabilities.
将冰冷的离别化为温暖的 Skill，欢迎加入数字生命1.0！Transforming cold farewells into warm skills? It's giving rebirth era. Welcome to Digital Life 1.0. 🫶
🕷️ An adaptive Web Scraping framework that handles everything from a single request to a full-scale crawl!
A feature-rich command-line audio/video downloader
Self-evolving Context Database for AI Agents. Unify Agent Memory, Knowledge RAG and Skills.
Open-source AI job search: scan job portals, evaluate listings with a structured A-F rubric into a 1.0-5.0 score, tailor your CV, track applications — runs locally in your AI coding CLI (Claude Code, Codex, OpenCode, Antigravity…)