arXiv:2504.18594v3 Announce Type: replace-cross Abstract: Compared to untargeted attacks, targeted transfer-based attack is still suffering from much lower Attack Success Rates (ASRs), although significant improvements have been achieved by kinds of methods, such as diversifying input, stabilizing the gradient, and re-training surrogate models. In this paper, we find that adversarial examples generated by existing methods rely heavily on a small subset of surrogate model parameters, which in turn limits their transferability to unseen target models. Inspired by this, we propose the Random Parameter Pruning Attack (RaPA), which introduces parameter-level randomization during the attack process. At each optimization step, RaPA randomly prunes model parameters to generate diverse yet semantically consistent surrogate variants.We show this parameter-level randomization is equivalent to adding an importance-equalization regularizer, thereby alleviating the over-reliance issue. Extensive experiments across both CNN and Transformer architectures demonstrate that RaPA substantially enhances transferability. In the challenging case of transferring from CNN-based to Transformer-based models, RaPA achieves up to 11.7% higher average ASRs than state-of-the-art baselines(with 33.3% ASRs), while being training-free, cross-architecture efficient, and easily integrated into existing attack frameworks. Code is available in https://github.com/molarsu/RaPA.

For years Mike McClary sold the Guardian LTE Flashlight, a heavy-duty black model, online through his small outdoor brand. The product, designed for brightness and durability, became one of his most popular items ever. Even after he stopped offering it around 2017, customers kept sending him emails asking where they could buy it. 

When McClary decided to revisit the Guardian flashlight in 2025, he didn’t begin the way he might have in the past, by combing through supplier listings and sending inquiries to factories. Instead, he opened Accio, an AI sourcing and researching tool on Alibaba.com.

For small entrepreneurs in the US, deciding what to sell and where to make it has traditionally been a slow, labor-intensive process that can take months. Now that work is increasingly being done by AI tools like Accio, which help connect businesses with manufacturers in countries including China and India. Business owners and e-commerce experts told MIT Technology Review that these AI tools are making sourcing more accessible and significantly shortening the time it takes to go from product idea to launch. 

McClary, 51, who runs his business from his Illinois living room, has sold products ranging from leather conditioner to camping lights, including one rechargeable lantern that brought in half a million dollars. Like many small online merchants, he built his business by being extremely scrappy—spotting demand for a product, tweaking existing designs, finding a factory, doing modest marketing, and getting the goods in front of customers fast. 

This time, though, he began by telling Accio about the flashlight’s original design, production cost, and profit margin. Then Accio suggested several changes, making it smaller and slightly less bright and switching its charging method to battery power. It also identified a manufacturer in Ningbo, China, that McClary said could cut the manufacturing cost from $17 to about $2.50 per unit.

McClary took the process from there, contacting the supplier himself to discuss the revised design. Within a month, the new version of the Guardian flashlight was back up for sale on Amazon and on his brand’s website.

The new factory hunt

Although Alibaba is better known for owning Taobao, the biggest shopping site in China, its first business was Alibaba.com, the primary website that lists Chinese factories open for bulk orders. Placing an order with a manufacturer usually requires far more than clicking “Buy.” Sellers often spend days or weeks browsing listings, comparing suppliers’ reviews and manufacturing capacities, asking about minimum order quantities, requesting samples, and negotiating timelines and customization options. 

But Accio has gained significant momentum by changing how that sourcing gets done. Launched in 2024, Accio exceeded 10 million monthly active users in March 2026, according to the company. That means about one in five Alibaba users consults with AI about product sourcing.

Accio’s interface looks a lot like ChatGPT or Claude: Users type a question into an empty box and choose between “fast” and “thinking” modes. But when asked about products, the tool returns more than text, offering charts, links, and visuals and asking follow-up questions to clarify the buyer’s needs. It then narrows the field to one or a handful of suppliers that appear capable of delivering. After that, the human work begins: Users still have to reach out to suppliers themselves and negotiate the details.

Zhang Kuo, the president of Alibaba.com, told MIT Technology Review that the tool is built on multiple frontier models, including the company’s own Qwen series, a popular family of open-source large language models. The system is able to pull from the site’s millions of supplier profiles and is trained on 26 years of proprietary transaction data.

For tasks like product research and sourcing analysis, the tool “blows it away” compared with general AI tools like ChatGPT, says Richard Kostick, CEO of the beauty brand 100% Pure.

Many websites have tried using AI to assist shopping, but Alibaba has been one of the most aggressive. In March, Eddie Wu, CEO of the site’s parent company Alibaba Group, told managers that integrating the company’s core services with Qwen’s AI capabilities is a top priority. During a Chinese New Year promotion of Qwen’s personal shopping AI agent, where the company gave away cash, customers placed 200 million orders, the firm says.

Vincenzo Toscano, an e-commerce seller and consultant, recommended Accio to his clients before deciding to try it himself for a new sunglasses brand. He came in with a rough vision: a brand shaped by his Italian heritage, his personal style, and a boutique aesthetic. He says the AI helped turn that concept into something more concrete, suggesting materials, refining the look, and pointing to design ideas that felt current.

But the tool has clear limits. McClary, who uses AI tools regularly, says Accio is strongest when it comes to product ideation, but less helpful on marketing questions such as advertising and social media outreach. To use it well, he says, buyers still need to challenge its recommendations, since some can be generic.

The rest of the business

As platforms become more AI-driven, manufacturers are adjusting too. Sally Li, a representative at a makeup packaging company in Wuhan, China, says her firm has started writing more detailed product descriptions and adding information about its equipment and manufacturing experience on Alibaba.com because it suspects those details make its listings more likely to be surfaced by AI.

Yan says manufacturers cannot tell whether an inquiry from a customer was generated or guided by AI, and that her firm is not using AI to negotiate pricing or product details.

“AI agents are increasingly used by people to assist purchase decisions and even directly making transactions, and with clear guardrails, they can become extremely useful,” says Jiaxin Pei, a research scientist at the Stanford Institute for Human-Centered AI, “but agents need to act transparently, securely, and in the customer’s best interest.” Pei says developers of these tools should disclose the data they collect and the incentives built into them to ensure that the marketplace remains fair.

Zhang, of Alibaba.com, says Accio currently does not include advertising. Suppliers can pay for higher placement in Alibaba.com’s regular search results, but Zhang says Accio is “not integrated” with that system. “We haven’t had a clear answer in terms of how to monetize this tool,” he says. For now, users can pay for additional tokens to continue chatting with the agent after their free queries run out.

Sellers say that while AI tools have made it easier to come up with ideas and get a business off the ground, they do not replace the core skills that make someone good at e-commerce. McClary believes that even when sellers have access to the same market information, some are still better at making decisions, acting quickly, and actually delivering on orders. Those differences, he says, still go a long way.

Toscano, the brand founder and e-commerce consultant, feels good about officially launching his new brand of sunglasses in just a few months: “We [small business owners] always have to bootstrap a lot of decisions. Deciding what to sell often comes down to an educated guess,” he says, “And we’re now in an era when making those decisions is easier than ever.”

arXiv:2604.01618v1 Announce Type: cross Abstract: Vision-language-action (VLA) models have shown strong performance in robotic manipulation, yet their robustness to physically realizable adversarial attacks remains underexplored. Existing studies reveal vulnerabilities through language perturbations and 2D visual attacks, but these attack surfaces are either less representative of real deployment or limited in physical realism. In contrast, adversarial 3D textures pose a more physically plausible and damaging threat, as they are naturally attached to manipulated objects and are easier to deploy in physical environments. Bringing adversarial 3D textures to VLA systems is nevertheless nontrivial. A central obstacle is that standard 3D simulators do not provide a differentiable optimization path from the VLA objective function back to object appearance, making it difficult to optimize through an end-to-end manner. To address this, we introduce Foreground-Background Decoupling (FBD), which enables differentiable texture optimization through dual-renderer alignment while preserving the original simulation environment. To further ensure that the attack remains effective across long-horizon and diverse viewpoints in the physical world, we propose Trajectory-Aware Adversarial Optimization (TAAO), which prioritizes behaviorally critical frames and stabilizes optimization with a vertex-based parameterization. Built on these designs, we present Tex3D, the first framework for end-to-end optimization of 3D adversarial textures directly within the VLA simulation environment. Experiments in both simulation and real-robot settings show that Tex3D significantly degrades VLA performance across multiple manipulation tasks, achieving task failure rates of up to 96.7\%. Our empirical results expose critical vulnerabilities of VLA systems to physically grounded 3D adversarial attacks and highlight the need for robustness-aware training.

arXiv:2604.02324v1 Announce Type: cross Abstract: Language models (LMs) are increasingly extended with new learnable vocabulary tokens for domain-specific tasks, such as Semantic-ID tokens in generative recommendation. The standard practice initializes these new tokens as the mean of existing vocabulary embeddings, then relies on supervised fine-tuning to learn their representations. We present a systematic analysis of this strategy: through spectral and geometric diagnostics, we show that mean initialization collapses all new tokens into a degenerate subspace, erasing inter-token distinctions that subsequent fine-tuning struggles to fully recover. These findings suggest that \emph{token initialization} is a key bottleneck when extending LMs with new vocabularies. Motivated by this diagnosis, we propose the \emph{Grounded Token Initialization Hypothesis}: linguistically grounding novel tokens in the pretrained embedding space before fine-tuning better enables the model to leverage its general-purpose knowledge for novel-token domains. We operationalize this hypothesis as GTI (Grounded Token Initialization), a lightweight grounding stage that, prior to fine-tuning, maps new tokens to distinct, semantically meaningful locations in the pretrained embedding space using only paired linguistic supervision. Despite its simplicity, GTI outperforms both mean initialization and existing auxiliary-task adaptation methods in the majority of evaluation settings across multiple generative recommendation benchmarks, including industry-scale and public datasets. Further analyses show that grounded embeddings produce richer inter-token structure that persists through fine-tuning, corroborating the hypothesis that initialization quality is a key bottleneck in vocabulary extension.

arXiv:2603.29291v1 Announce Type: cross Abstract: Composed Image Retrieval (CIR) uses a reference image and a modification text as a query to retrieve a target image satisfying the requirement of ``modifying the reference image according to the text instructions''. However, existing CIR methods face two limitations: (1) frequency bias leading to ``Rare Sample Neglect'', and (2) susceptibility of similarity scores to interference from hard negative samples and noise. To address these limitations, we confront two key challenges: asymmetric rare semantic localization and robust similarity estimation under hard negative samples. To solve these challenges, we propose the Modification frEquentation-rarity baLance neTwork MELT. MELT assigns increased attention to rare modification semantics in multimodal contexts while applying diffusion-based denoising to hard negative samples with high similarity scores, enhancing multimodal fusion and matching. Extensive experiments on two CIR benchmarks validate the superior performance of MELT. Codes are available at https://github.com/luckylittlezhi/MELT.

arXiv:2603.29292v1 Announce Type: cross Abstract: Improving the code generation capabilities of large language models (LLMs) typically relies on supervised fine-tuning or preference optimization, both of which require costly external resources such as powerful teacher models or reliable test units. However, in real-world scenarios, it is much harder to obtain reference solutions and test oracles than problem descriptions and test inputs. In this paper, we tackle a challenging yet realistic question: Can a code language model improve itself without access to a superior teacher and a test oracle? To answer this, we propose ConSelf, a self-improving approach built upon two key ideas. First, we introduce code semantic entropy, a novel metric that measures problem-level uncertainty by assessing the functional diversity of program behaviors, enabling a curriculum construction with the most learnable problems. Second, we present consensus-driven direct preference optimization (Con-DPO), a preference-based fine-tuning method that weights each preference pair by its behavioral consensus, thereby mitigating the impact of noisy self-generated supervision. Experiments on various benchmarks and backbone LLMs demonstrate that ConSelf significantly outperforms baselines, validating the effectiveness of semantic entropy-based curriculum construction and consensus-driven optimization in improving code generation without external supervision.

arXiv:2603.12953v1 Announce Type: cross Abstract: Neuro-symbolic reasoning increasingly demands frameworks that unite the formal rigor of logic with the interpretability of large language models (LLMs). We introduce an end to end explainability by construction pipeline integrating the Automated Theorem Generator Delta1 based on the full triangular standard contradiction (FTSC) with LLMs. Delta1 deterministically constructs minimal unsatisfiable clause sets and complete theorems in polynomial time, ensuring both soundness and minimality by construction. The LLM layer verbalizes each theorem and proof trace into coherent natural language explanations and actionable insights. Empirical studies across health care, compliance, and regulatory domains show that Delta1 and LLM enables interpretable, auditable, and domain aligned reasoning. This work advances the convergence of logic, language, and learning, positioning constructive theorem generation as a principled foundation for neuro-symbolic explainable AI.

arXiv:2603.08605v3 Announce Type: replace-cross Abstract: Background and objectives: Colorectal cancer histopathological grading depends on accurate segmentation of glandular structures. Current deep learning approaches rely on large scale pixel level annotations that are labor intensive and difficult to obtain in routine clinical practice. Weakly supervised semantic segmentation offers a promising alternative. However, class activation map based methods often produce incomplete pseudo masks that emphasize highly discriminative regions and fail to supervise unannotated glandular structures. We propose a weakly supervised teacher student framework that leverages sparse pathologist annotations and an Exponential Moving Average stabilized teacher network to generate refined pseudo masks. Methods: The framework integrates confidence based filtering, adaptive fusion of teacher predictions with limited ground truth, and curriculum guided refinement to progressively segment unannotated glandular regions. The method was evaluated on an institutional colorectal cancer cohort from The Ohio State University Wexner Medical Center consisting of 60 hematoxylin and eosin stained whole slide images and on public datasets including the Gland Segmentation dataset, TCGA COAD, TCGA READ, and SPIDER. Results: On the Gland Segmentation dataset the framework achieved a mean Intersection over Union of 80.10 and a mean Dice coefficient of 89.10. Cross cohort evaluation demonstrated robust generalization on TCGA COAD and TCGA READ without additional annotations, while reduced performance on SPIDER reflected domain shift. Conclusions: The proposed framework provides an annotation efficient and generalizable approach for gland segmentation in colorectal histopathology.

Feng Qingyang had always hoped to launch his own company, but he never thought this would be how—or that the day would come this fast. 

Feng, a 27-year-old software engineer based in Beijing, started tinkering with OpenClaw, a popular new open-source AI tool that can take over a device and autonomously complete tasks for a user, in January. He was immediately hooked, and before long he was helping other curious tech workers with less technical proficiency install the AI agent.

Feng soon realized this could be a lucrative opportunity. By the end of January, he had set up a page on Xianyu, a secondhand shopping site, advertising “OpenClaw installation support.” “No need to know coding or complex terms. Fully remote,” reads the posting. “Anyone can quickly own an AI assistant, available within 30 minutes.” 

At the same time, the broader Chinese public was beginning to catch on—and the tool, which had begun as a niche interest among tech workers, started to evolve into a popular sensation.

Feng quickly became inundated with requests, and he started chatting with customers and managing orders late into the night. At the end of February, he quit his job. His side gig has now grown into a full-fledged professional operation with over 100 employees. So far, the store has handled 7,000 orders, each worth about 248 RMB or approximately $34. 

“Opportunities are always fleeting,” says Feng. “As programmers, we are the first to feel the winds shift.”

Feng is among a small cohort of savvy early adopters turning China’s OpenClaw craze into cash. As users with little technical background want in, a cottage industry of people offering installation services and preconfigured hardware has sprung up to meet them. The sudden rise of these tinkerers and impromptu consultants shows just how eager the general public in China is to adopt cutting-edge AI—even when there are huge security risks. 

A “lobster craze”

“Have you raised a lobster yet?” 

Xie Manrui, a 36-year-old software engineer in Shenzhen, says he has heard this question nonstop over the past month. “Lobster” is the nickname Chinese users have given to OpenClaw—a reference to its logo.

Xie, like Feng, has been experimenting with OpenClaw since January. He’s built new open-source tools on top of the ecosystem, including one that visualizes the agent’s progress as an animated little desktop worker and another that lets users voice-chat with it. 

“I’ve met so many new people through ‘lobster raising,’” says Xie. “Many are lawyers or doctors, with little technical background, but all dedicated to learning new things.”

Lobsters are indeed popping up everywhere in China right now—on and offline. In February, for instance, the entrepreneur and tech influencer Fu Sheng hosted a livestream showing off OpenClaw’s capabilities that got 20,000 views. And just last weekend, Xie attended three different OpenClaw events in Shenzhen, each drawing more than 500 people. These self-organized, unofficial gatherings feature power users, influencers, and sometimes venture capitalists as speakers. The biggest event Xie attended, on March 7, drew more than 1,000 people; in the packed venue, he says, people were shoulder to shoulder, with many attendees unable to even get a seat.

Now China’s AI giants are starting to piggyback on the trend too, promoting their models, APIs,  and cloud services (which can be used with OpenClaw), as well as their own OpenClaw-like agents. Earlier this month, Tencent held a public event offering free installation support for OpenClaw, drawing long lines of people waiting for help, including elderly users and children.

This sudden burst in popularity has even prompted local governments to get involved. Earlier this month the government of Longgang, a district in Shenzhen, released several policies to support OpenClaw-related ventures, including free computing credits and cash rewards for standout projects. Other cities, including Wuxi, have begun rolling out similar measures.

These policies only catalyze what’s already in the air. “It was not until my father, who is 77, asked me to help install a ‘lobster’ for him that I realized this thing is truly viral,” says Henry Li, a software engineer based in Beijing. 

A programmer gold rush

What’s making this moment particularly lucrative for people with technical skills, like Feng, is that so many people want OpenClaw, but not nearly as many have the capabilities to access it. Setting it up requires a level of technical knowledge most people do not possess, from typing commands into a black terminal window to navigating unfamiliar developer platforms. On the hardware side, an older or budget laptop may struggle to run it smoothly. And if the tool is not installed on a device separate from someone’s everyday computer, or if the data accessible to OpenClaw is not properly partitioned, the user’s privacy could be at risk—opening the door to data leaks and even malicious attacks. 

Chris Zhao, known as “Qi Shifu” online, organizes OpenClaw social media groups and events in Beijing. On apps like Rednote and Jike, Zhao routinely shares his thoughts on AI, and he asks other interested users to leave their WeChat ID so he can invite them to a semi-private group chat. The proof required to join is a screenshot that shows your “lobster” up and running. Zhao says that even in group chats for experienced users, hardware and cloud setup remain a constant topic of discussion.

The relatively high bar for setting up OpenClaw has generated a sense of exclusivity, creating a natural opening for a service industry to start unfolding around it. On Chinese e-commerce platforms like Taobao and JD, a simple search for “OpenClaw” now returns hundreds of listings, most of them installation guides and technical support packages aimed at nontechnical users, priced anywhere from 100 to 700 RMB (approximately $15 to $100). At the higher end, many vendors offer to come to help you in person. 

Like Feng, most providers of these services are early adopters with some technical ability who are looking for a side gig. But as demand has surged, some have found themselves overwhelmed. Xie, the developer in Shenzhen who created tools to layer on OpenClaw, was asked by a friend who runs one such business to help out over the weekend; the friend had a customer who worked in e-commerce and had little technical experience, so Xie had to show up in person to get it done. He walked away with 600 RMB ($87) for the afternoon.

The growing demand has also pushed vendors like Feng to expand quickly. He has now standardized his operation into tiers: a basic installation, a custom package where users can make specific requests like configuring a preferred chat app, and an ongoing tutoring service for those who want a hand to hold as they find their footing with the technology.

Other vendors in China are making money combining OpenClaw with hardware. Li Gong, a Shenzhen-based seller of refurbished Mac computers, was among the first online sellers to do this—offering Mac minis and MacBooks with OpenClaw preinstalled. Because OpenClaw is designed to operate with deep access to a hard drive and can run continuously in the background unattended, many users prefer to install it on a separate device rather than on the one they use every day. This would help prevent bad actors from infiltrating the program and immediately gaining access to a wide swathe of someone’s personal information. Many turn to secondhand or refurbished options to keep the cost down. Li says that in the last two weeks, orders have increased eightfold.

Though OpenClaw itself is a new technology, the general practice of buying software bundles, downloading third-party packages, and seeking out modified devices is nothing new for many Chinese internet users, says Tianyu Fang, a PhD candidate studying the history of technology at Harvard University. Many users pay for one-off IT support services for tasks from installing Adobe software to jailbreaking a Kindle.

Still, not everyone is getting swept up. Jiang Yunhui, a tech worker based in Ningbo, worries that ordinary users who struggle with setup may not be the right audience for a technology that is still effectively in testing. 

“The hype in first-tier cities can be a little overblown,” he says. “The agent is still a proof of concept, and I doubt it would be of any life-changing use to the average person for now.” He argues that using it safely and getting anything meaningful out of it requires a level of technical fluency and independent judgment that most new users simply don’t have yet.

He’s not alone in his concerns. On March 10, the Chinese cybersecurity regulator CNCERT issued a warning about the security and data risks tied to OpenClaw, saying it heightens users’ exposure to data breaches.

Despite the potential pitfalls, though, China’s enthusiasm for OpenClaw doesn’t seem to be slowing.

Feng, now flush with the earnings from his operation, wants to use the momentum—and the capital—to keep building out his own venture with AI tools at the center of it.

“With OpenClaw and other AI agents, I want to see if I can run a one-person company,” he says. “I’m giving myself one year.”

arXiv:2603.07648v1 Announce Type: cross Abstract: Recent advances in Visual-Language-Action (VLA) models have shown promising potential for robotic manipulation tasks. However, real-world robotic tasks often involve long-horizon, multi-step problem-solving and require generalization for continual skill acquisition, extending beyond single actions or skills. These challenges present significant barriers for existing VLA models, which use monolithic action decoders trained on aggregated data, resulting in poor scalability. To address these challenges, we propose AtomicVLA, a unified planning-and-execution framework that jointly generates task-level plans, atomic skill abstractions, and fine-grained actions. AtomicVLA constructs a scalable atomic skill library through a Skill-Guided Mixture-of-Experts (SG-MoE), where each expert specializes in mastering generic yet precise atomic skills. Furthermore, we introduce a flexible routing encoder that automatically assigns dedicated atomic experts to new skills, enabling continual learning. We validate our approach through extensive experiments. In simulation, AtomicVLA outperforms $\pi_{0}$ by 2.4\% on LIBERO, 10\% on LIBERO-LONG, and outperforms $\pi_{0}$ and $\pi_{0.5}$ by 0.22 and 0.25 in average task length on CALVIN. Additionally, our AtomicVLA consistently surpasses baselines by 18.3\% and 21\% in real-world long-horizon tasks and continual learning. These results highlight the effectiveness of atomic skill abstraction and dynamic expert composition for long-horizon and lifelong robotic tasks. The project page is \href{https://zhanglk9.github.io/atomicvla-web/}{here}.

arXiv:2603.08605v1 Announce Type: cross Abstract: Background and objectives: Colorectal cancer histopathological grading depends on accurate segmentation of glandular structures. Current deep learning approaches rely on large scale pixel level annotations that are labor intensive and difficult to obtain in routine clinical practice. Weakly supervised semantic segmentation offers a promising alternative. However, class activation map based methods often produce incomplete pseudo masks that emphasize highly discriminative regions and fail to supervise unannotated glandular structures. We propose a weakly supervised teacher student framework that leverages sparse pathologist annotations and an Exponential Moving Average stabilized teacher network to generate refined pseudo masks. Methods: The framework integrates confidence based filtering, adaptive fusion of teacher predictions with limited ground truth, and curriculum guided refinement to progressively segment unannotated glandular regions. The method was evaluated on an institutional colorectal cancer cohort from The Ohio State University Wexner Medical Center consisting of 60 hematoxylin and eosin stained whole slide images and on public datasets including the Gland Segmentation dataset, TCGA COAD, TCGA READ, and SPIDER. Results: On the Gland Segmentation dataset the framework achieved a mean Intersection over Union of 80.10 and a mean Dice coefficient of 89.10. Cross cohort evaluation demonstrated robust generalization on TCGA COAD and TCGA READ without additional annotations, while reduced performance on SPIDER reflected domain shift. Conclusions: The proposed framework provides an annotation efficient and generalizable approach for gland segmentation in colorectal histopathology.

arXiv:2509.08612v3 Announce Type: replace-cross Abstract: Aspect-based sentiment analysis (ABSA) aims to identify aspect terms and determine their sentiment polarity. While dependency trees combined with contextual semantics provide structural cues, existing approaches often rely on dot-product similarity and fixed graphs, which limit their ability to capture nonlinear associations and adapt to noisy contexts. To address these limitations, we propose the Optimal Transport-Enhanced Syntactic-Semantic Graph Network (OTESGN), a model that jointly integrates structural and distributional signals. Specifically, a Syntactic Graph-Aware Attention module models global dependencies with syntax-guided masking, while a Semantic Optimal Transport Attention module formulates aspect-opinion association as a distribution matching problem solved via the Sinkhorn algorithm. An Adaptive Attention Fusion mechanism balances heterogeneous features, and contrastive regularization enhances robustness. Extensive experiments on three benchmark datasets (Rest14, Laptop14, and Twitter) demonstrate that OTESGN delivers state-of-the-art performance. Notably, it surpasses competitive baselines by up to +1.30 Macro-F1 on Laptop14 and +1.01 on Twitter. Ablation studies and visualization analyses further highlight OTESGN's ability to capture fine-grained sentiment associations and suppress noise from irrelevant context.

arXiv:2603.03761v1 Announce Type: new Abstract: LLM agents are rapidly becoming the practical interface for task automation, yet the ecosystem lacks a principled way to choose among an exploding space of deployable configurations. Existing LLM leaderboards and tool/agent benchmarks evaluate components in isolation and remain fragmented across tasks, metrics, and candidate pools, leaving a critical research gap: there is little query-conditioned supervision for learning to recommend end-to-end agent configurations that couple a backbone model with a toolkit. We address this gap with AgentSelect, a benchmark that reframes agent selection as narrative query-to-agent recommendation over capability profiles and systematically converts heterogeneous evaluation artifacts into unified, positive-only interaction data. AgentSelectcomprises 111,179 queries, 107,721 deployable agents, and 251,103 interaction records aggregated from 40+ sources, spanning LLM-only, toolkit-only, and compositional agents. Our analyses reveal a regime shift from dense head reuse to long-tail, near one-off supervision, where popularity-based CF/GNN methods become fragile and content-aware capability matching is essential. We further show that Part~III synthesized compositional interactions are learnable, induce capability-sensitive behavior under controlled counterfactual edits, and improve coverage over realistic compositions; models trained on AgentSelect also transfer to a public agent marketplace (MuleRun), yielding consistent gains on an unseen catalog. Overall, AgentSelect provides the first unified data and evaluation infrastructure for agent recommendation, which establishes a reproducible foundation to study and accelerate the emerging agent ecosystem.

arXiv:2603.03415v1 Announce Type: cross Abstract: In this work, we investigate how Large Language Models (LLMs) adapt their internal representations when encountering inputs of increasing difficulty, quantified as the degree of out-of-distribution (OOD) shift. We reveal a consistent and quantifiable phenomenon: as task difficulty increases, whether through harder reasoning questions, longer contexts, or adding answer choices, the last hidden states of LLMs become substantially sparser. In short, \textbf{\textit{the farther the shift, the sparser the representations}}. This sparsity--difficulty relation is observable across diverse models and domains, suggesting that language models respond to unfamiliar or complex inputs by concentrating computation into specialized subspaces in the last hidden state. Through a series of controlled analyses with a learning dynamic explanation, we demonstrate that this sparsity is not incidental but an adaptive mechanism for stabilizing reasoning under OOD. Leveraging this insight, we design \textit{Sparsity-Guided Curriculum In-Context Learning (SG-ICL)}, a strategy that explicitly uses representation sparsity to schedule few-shot demonstrations, leading to considerable performance enhancements. Our study provides new mechanistic insights into how LLMs internalize OOD challenges. The source code is available at the URL: https://github.com/MingyuJ666/sparsityLLM.

arXiv:2603.03782v1 Announce Type: cross Abstract: Shared-account usage is common on streaming and e-commerce platforms, where multiple users share one account. Existing shared-account sequential recommendation (SSR) methods often assume a fixed number of latent users per account, limiting their ability to adapt to diverse sharing patterns and reducing recommendation accuracy. Recent latent reasoning technique applied in sequential recommendation (SR) generate intermediate embeddings from the user embedding (e.g, last item embedding) to uncover users' potential interests, which inspires us to treat the problem of inferring the number of latent users as generating a series of intermediate embeddings, shifting from inferring preferences behind user to inferring the users behind account. However, the last item cannot be directly used for reasoning in SSR, as it can only represent the behavior of the most recent latent user, rather than the collective behavior of the entire account. To address this, we propose DisenReason, a two-stage reasoning method tailored to SSR. DisenReason combines behavior disentanglement stage from frequency-domain perspective to create a collective and unified account behavior representation, which serves as a pivot for latent user reasoning stage to infer the number of users behind the account. Experiments on four benchmark datasets show that DisenReason consistently outperforms all state-of-the-art baselines across four benchmark datasets, achieving relative improvements of up to 12.56\% in MRR@5 and 6.06\% in Recall@20.

arXiv:2603.02697v1 Announce Type: cross Abstract: This paper presents ShareVerse, a video generation framework enabling multi-agent shared world modeling, addressing the gap in existing works that lack support for unified shared world construction with multi-agent interaction. ShareVerse leverages the generation capability of large video models and integrates three key innovations: 1) A dataset for large-scale multi-agent interactive world modeling is built on the CARLA simulation platform, featuring diverse scenes, weather conditions, and interactive trajectories with paired multi-view videos (front/ rear/ left/ right views per agent) and camera data. 2) We propose a spatial concatenation strategy for four-view videos of independent agents to model a broader environment and to ensure internal multi-view geometric consistency. 3) We integrate cross-agent attention blocks into the pretrained video model, which enable interactive transmission of spatial-temporal information across agents, guaranteeing shared world consistency in overlapping regions and reasonable generation in non-overlapping regions. ShareVerse, which supports 49-frame large-scale video generation, accurately perceives the position of dynamic agents and achieves consistent shared world modeling.

arXiv:2509.22613v2 Announce Type: replace Abstract: Recent reinforcement learning (RL) methods have substantially enhanced the planning capabilities of Large Language Models (LLMs), yet the theoretical basis for their effectiveness remains elusive. In this work, we investigate RL's benefits and limitations through a tractable graph-based abstraction, focusing on policy gradient (PG) and Q-learning methods. Our theoretical analyses reveal that supervised fine-tuning (SFT) may introduce co-occurrence-based spurious solutions, whereas RL achieves correct planning primarily through exploration, underscoring exploration's role in enabling better generalization. However, we also show that PG suffers from diversity collapse, where output diversity decreases during training and persists even after perfect accuracy is attained. By contrast, Q-learning provides two key advantages: off-policy learning and diversity preservation at convergence. We further demonstrate that careful reward design is necessary to prevent Q-value bias in Q-learning. Finally, applying our framework to the real-world planning benchmark Blocksworld, we confirm that these behaviors manifest in practice.

arXiv:2506.01502v3 Announce Type: replace-cross Abstract: Learning population dynamics involves recovering the underlying process that governs particle evolution, given evolutionary snapshots of samples at discrete time points. Recent methods frame this as an energy minimization problem in probability space and leverage the celebrated JKO scheme for efficient time discretization. In this work, we introduce $\texttt{iJKOnet}$, an approach that combines the JKO framework with inverse optimization techniques to learn population dynamics. Our method relies on a conventional $\textit{end-to-end}$ adversarial training procedure and does not require restrictive architectural choices, e.g., input-convex neural networks. We establish theoretical guarantees for our methodology and demonstrate improved performance over prior JKO-based methods. The code of $\texttt{iJKOnet}$ is available at https://github.com/MuXauJl11110/iJKOnet.

arXiv:2602.18493v1 Announce Type: cross Abstract: Long-context LLMs and Retrieval-Augmented Generation (RAG) systems process information passively, deferring state tracking, contradiction resolution, and evidence aggregation to query time, which becomes brittle under ultra long streams with frequent updates. We propose the Unified Memory Agent (UMA), an end-to-end reinforcement learning framework that unifies memory operations and question answering within a single policy. UMA maintains a dual memory representation: a compact core summary for global context and a structured Memory Bank that supports explicit CRUD (create, update, delete, reorganize) over key value entries, enabling proactive consolidation during streaming. To evaluate long-horizon memory behavior, we introduce Ledger-QA, a diagnostic benchmark for continuous state tracking where answers are latent values derived from accumulated updates rather than lo cal span retrieval. Across 13 datasets spanning Ledger-QA, Test-Time Learning, and Accurate Retrieval, UMA substantially outperforms long-context and RAG baselines on dynamic reasoning and learning tasks while remaining competitive on standard retrieval benchmarks, underscoring the importance of learned, end-to-end memory management.

arXiv:2602.19113v1 Announce Type: cross Abstract: Spatio-temporal forecasting is fundamental to intelligent systems in transportation, climate science, and urban planning. However, training deep learning models on the massive, often redundant, datasets from these domains presents a significant computational bottleneck. Existing solutions typically focus on optimizing model architectures or optimizers, while overlooking the inherent inefficiency of the training data itself. This conventional approach of iterating over the entire static dataset each epoch wastes considerable resources on easy-to-learn or repetitive samples. In this paper, we explore a novel training-efficiency techniques, namely learning from complexity with dynamic sample pruning, ST-Prune, for spatio-temporal forecasting. Through dynamic sample pruning, we aim to intelligently identify the most informative samples based on the model's real-time learning state, thereby accelerating convergence and improving training efficiency. Extensive experiments conducted on real-world spatio-temporal datasets show that ST-Prune significantly accelerates the training speed while maintaining or even improving the model performance, and it also has scalability and universality.