Daily AI Papers — August 15, 2026

1. Beyond Starry Night: Shortcut-Aware Control-State Planning for Artist-Grounded Text to Image Generation

Authors: Kuan Xing, Ye Wang, Changyi Gan, Yuheng Li, Thao Nguyen, Yi Chang, Yilin Wang arXiv: arxiv.org/abs/2608.06751 Summary: Artist-grounded image generation requires more than appending an artist name to a prompt. Image models often respond to artist names through canonical shortcuts, such as recurring motifs, generic palettes, or overrepresented period signatures, rather than preserving the user’s intended scene. Trending because: 27 HuggingFace upvotes + among the more-upvoted papers in this weekend’s feed.

2. Full-bandwidth transformer

Authors: Xi Wang, Ziyang Cai, Zheng Zhan, Harry Dong, Ying Fan, Gustavo de Rosa, Tim Pearce, John Langford arXiv: arxiv.org/abs/2608.08888 Summary: Autoregressive transformers compute along two axes: horizontally across generated tokens, and vertically through model depth. Dense attention gives each token broad horizontal access to the past, but the vertical feedback channel between decoding steps remains narrow: only the sampled token returns to the bottom of the stack, while the top-layer hidden state is discarded. Trending because: 10 HuggingFace upvotes + among the more-upvoted papers in this weekend’s feed.

3. Intent Speaks Louder: Controllable User Simulation Beyond Response Imitation

Authors: Bo Wang, Ruixing Zhang, Yunqi Liu, Yang Zhang, Liangzhe Han, Tongyu Zhu, Leilei Sun arXiv: arxiv.org/abs/2608.09420 Summary: User simulators are widely used as scalable environments for training and evaluating interactive assistants. Generating the next user turn is inherently one-to-many: the same profile and dialogue context may support multiple plausible continuations with different local interaction intents. Trending because: 8 HuggingFace upvotes + among the more-upvoted papers in this weekend’s feed.

4. An AI4AI Framework for Visual Token Pruning

Authors: Zhen Liu, Wenli Huang, Wei Song, Yuhan Liu, Zhiqin Yang, Jingwen Fu arXiv: arxiv.org/abs/2608.07193 Summary: Visual-token pruning can substantially reduce the inference cost of multimodal large language models (MLLMs), yet existing methods largely rely on fixed, handcrafted heuristics and costly expert trial and error. As pruning objectives, budgets, and model architectures diversify, manually navigating the expanding design space becomes increasingly difficult. Trending because: 7 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

5. Don’t Scroll Back: Missing-Evidence Memory for Streaming Dialogue Summarization

Authors: Hyangsuk Min, Hwanjun Song arXiv: arxiv.org/abs/2608.09043 Summary: Users of modern platforms repeatedly need summaries of recent dialogue, but the window rarely contains enough context to be interpreted on its own. We formalize this setting as streaming dialogue summarization, where a system must summarize a current window using selective memory from an unbounded history under a fixed budget. Trending because: 7 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

6. Gaming Without an Attacker: Benchmark Fingerprinting in LLM-Driven Search Under Selection Pressure

Authors: Víctor Gallego arXiv: arxiv.org/abs/2608.08722 Summary: Benchmarks for systems that are optimized against the evaluation signal measure something different from what they claim. We document this concretely in two GPU-kernel-optimization suites with held-out generalization gates: Metal-Sci (10 scientific-compute tasks) and Metal-ZK (12 zero-knowledge/cryptographic tasks), in which three frontier LLMs (Opus 4.7, Gemini 3.1 Pro, GPT-5.5) propose Metal kernels inside a (1{+}1) evolutionary loop with rich feedback. Trending because: 7 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

7. Vision-Language Grounding as Bidirectional Concept Correspondence

Authors: Jieyu Zhang, Ziqi Gao, Luke Zettlemoyer, Ranjay Krishna arXiv: arxiv.org/abs/2608.07886 Summary: Vision-language grounding connects language to visual content, yet most existing formulations reduce grounding to a unidirectional localization problem: given a prespecified text phrase or category name, identify the corresponding image region. This setup assumes that the relevant linguistic unit is already known, overlooking a more basic challenge in grounded communication: determining which parts of the text are visually referential and how they correspond to entities in the image. Trending because: 7 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

8. AVA-Encoder: Towards Agent-Native Video Representation Learning

Authors: Chuyue Li, Jinpeng Yu, Haozhe Wang, Tian Xueyun, Zhijing Zhang, Bingnan Li, Shuqi Gu, Kan Ren, Jiaming Liu, Ruihua Hua arXiv: arxiv.org/abs/2608.12313 Summary: Creative agents still lack an effective way to learn from high-quality human films, limiting their ability to produce cinematic-grade videos. A key challenge is the absence of a structured video representation that is both faithful to film content and directly usable for agentic reasoning and manipulation. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

9. Beyond Sequence Order: Syntax-Informed Positional Embeddings for Transformers

Authors: Haris Riaz, Hyungji Kim, Mihai Surdeanu arXiv: arxiv.org/abs/2608.06111 Summary: Positional embeddings (PE) in Transformers encode token distance and order but are largely agnostic to syntactic structure. We introduce Syntax-informed Positional Embeddings (SiPE), which learns a lightweight syntactic prior from dependency parses during pretraining and injects it across all three dominant PE families (absolute, relative, rotary), for both encoders and decoders, leaving self-attention and the rest of the architecture untouched. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

10. CEAA: A Cognitive Embodied Agents Architecture for Interactive Computing Systems

Authors: Aimilios Hadjiliasi, Louis Nisiotis arXiv: arxiv.org/abs/2608.09848 Summary: The development of embodied Intelligent Virtual Agents (IVAs) that have cognitive capabilities in real-time interactive virtual environments remains a challenge, even with today’s advancements in technology. Existing architectures are often focused on either the implementation of low-level reactive control systems that are constrained by commercial game engines, or high-level representations of reasoning models that can be difficult to implement in virtual worlds. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

11. Cultivar: A Contrastive and Locale-Oriented Translation Benchmark for Investigating Contamination and Localisation Robustness

Authors: Pinzhen Chen, Koel Dutta Chowdhury, Xiaoya Xu, David Tan, Doreen Osmelak, Ona de Gibert, Ariun-Erdene Tumurchuluun, Ashok Urlana, Fedor Sizov, Hale Sirin, Jesujoba Alabi, Karrar Talib Abed, Mateusz Klimaszewski, Nikolay Bogoychev, Niyati Bafna, Patricia Schmidtova, Preksha Manjunath Shanbhag, Sherrie Shen, Vilem Zouhar, Vivek Iyer, Yasser Hamidullah, Yusser Al Ghussin, Zheng Zhao arXiv: arxiv.org/abs/2608.09766 Summary: Multilingual translation benchmarks are typically sourced in English and translated into other languages, treating language pairs as the unit of evaluation—a design that is prone to contamination over time and overlooks locale and cultural considerations. We therefore advocate for source-contrastive evaluation and instantiate it with Cultivar, a localised subset of FLORES, which enables locale-specific translation evaluation. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

12. H2R-Bench: Benchmarking Human-to-Robot Manipulation Video Generation in World Models

Authors: Dingyi Rong, Yue Shi, Chaofan Ma, Jiezhang Cao, Zongrui Wang, Zeyu Zhang, Yao Mu, Guangtao Zhai, Ning Liu arXiv: arxiv.org/abs/2608.13049 Summary: Large-scale manipulation data is essential for robot learning, yet collecting robot demonstrations remains expensive and difficult to scale. Meanwhile, abundant egocentric human manipulation videos provide rich behavioral experiences, but transferring them across embodiments remains challenging due to differences between human hands and robotic end-effectors. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

13. MBA: Multimodal Benchmark and Agents for Real-World Business Ideation

Authors: Hojun Choi, Jaeyo Shin, Suin Lee, Hyunjung Shim arXiv: arxiv.org/abs/2608.11616 Summary: Agentic systems powered by large language models (LLMs) have opened new opportunities for business ideation. Yet existing approaches remain confined to a text-only paradigm, despite the inherently multimodal nature of real-world contexts. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

14. MMOOC: A Comprehensive Benchmark for Out-of-Context Evaluation in Multimodal Large Language Models

Authors: Wenjie Zhu, Yabin Zhang, Wenjun Zeng, Lei Zhang arXiv: arxiv.org/abs/2607.27637 Summary: Multimodal Large Language Models (MLLMs) have achieved strong performance on a wide range of vision-language tasks, but often fail under imperfect or shifted contexts. A reliable MLLM should refuse truly out-of-context (OOC) questions with subject-level context shifts while still answering shifted in-context (Shifted IC) questions with non-subject context shifts. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

15. MirrorWorld: Taming Video Diffusion Models for Mirror Reflection Generation

Authors: Youjun Zhao, Alex Warren, Gary K. L. Tam, Rynson W. H. Lau arXiv: arxiv.org/abs/2608.07463 Summary: Recent advances in video diffusion models (VDMs) have enabled high-fidelity video synthesis. However, generating mirror reflections remains challenging because the content within a mirror must remain consistent with the surrounding scene. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

16. OmniScientist: An Omni-Modal Omni-Discipline AI Scientist

Authors: Bobo Li, Hao Fei, Tianjie Ju, Mong-Li Lee, Wynne Hsu arXiv: arxiv.org/abs/2608.13558 Summary: Recent advances in foundation models have enabled AI scientists to automate increasingly complete research workflows, from hypothesis generation and code execution to manuscript preparation. Yet workflow coverage alone does not provide access to the full evidence on which scientific discovery depends. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

17. Parameter Exploration for RLVR via Variational Learning

Authors: Vatsal Venkatkrishna, Nico Daheim, Iryna Gurevych arXiv: arxiv.org/abs/2608.09805 Summary: Exploration has been a focus of reinforcement learning research for a long time. Recently, there has been growing evidence that it is also an important ingredient in LLM reinforcement learning recipes that can significantly impact downstream performance. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

18. Persistent Recursive Worlds Enable Autonomous Software Evolution

Authors: Beichen Huang, Zhenyu Liang, Bowen Zheng, Ran Cheng arXiv: arxiv.org/abs/2608.10450 Summary: Complex software systems develop over timescales that exceed the lifespan of any individual coding agent. Most agentic software systems preserve continuity through persistent sessions, memories, managers or shared context. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

19. SymDiag: Explainable Diagnosis for LLM Reasoning via Neuro-Symbolic Verification

Authors: Wenyao Cui, Huaping Zhang, Yongyi Huang, Qiuchi Li, Jian Xu, Cheng-Lin Liu, Chunxiao Gao, Juan Wang, Baohua Zhang arXiv: arxiv.org/abs/2608.08786 Summary: Large language models (LLMs) increasingly serve as data-driven reasoners, yet their chains-of-thought (CoT) can be unfaithful even when final answers are correct. Most existing ``verification’’ signals are not diagnostic: answer matching observes only the outcome, LLM-as-judge provides subjective and non-verifiable critiques, and scalar rewards (e.g., PRMs/RMs) offer little insight into where a multi-step derivation fails.We propose SymDiag, a neuro-symbolic framework that reframes reasoning verification as structured failure diagnosis. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.

20. Thought-Level Beam Search for Reasoning

Authors: Lijie Yang, Hongyin Luo, Jiawei Zhao, Tri Dao, Ravi Netravali arXiv: arxiv.org/abs/2608.08020 Summary: Test-time compute scaling is a primary driver of performance in large reasoning models (LRMs), but extreme inefficiency bounds current approaches, shifting the critical question from how much compute to spend, to where to allocate it. We formalize test-time reasoning as a constrained compute allocation problem over partial trajectories. Trending because: 6 HuggingFace upvotes + notable engagement given the weekend’s lighter activity.