Meta

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Learn how to fine-tune and evaluate LLMs with LangSmith for dataset management. Complete guide covers LLaMA2 and GPT-3.5 fine-tuning with practical examples.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Training and serving frontier models is now a networking problem as much as a compute problem. Collective operations like all-reduce and all-to-all synchronize thousands of accelerators during training, and the slowest transfer sets the pace for the entire job. Even small amounts of network friction directly strand significant compute capacity. This week, Meta introduced MetaRoCE. […] The post Meta AI Introduces MetaRoCE: A Clean-Sheet RDMA Transport Built for AI-Scale Ethernet appeared first on MarkTechPost.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：I have a complicated relationship with Hacker News. The site is the most important aggregator of geek news and a major source of traffic to this blog. At the same time, it has a fair number of toxic commenters, making i…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Open table formats are metadata layers that sit on top of data files in object storage, adding ACID transactions...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.21415v1 Announce Type: new Abstract: Large Vision-Language Models (LVLMs) have achieved remarkable performance across a wide range of tasks; however, they often inherit social biases from their training data, resulting in biased behavior when processing portraits from different social groups. Existing debiasing approaches typically compare token probabilities between the original and biased generations during decoding, but they are fundamentally limited by their reliance on a single, stereotyped viewpoint and fail to account for the diversity of social perspectives. Inspired by the social science principle that diversity fosters fairness, we propose Counterfactual Ensemble Decoding (CED), a novel framework that constructs multi-group counterfactual perspectives within the visual representation space and integrates them during decoding to promote equitable model behavior. CED first performs counterfactual steering in the visual space by identifying semantic directions associated with each social group and generating counterfactual representations along these directions, thereby offering diverse perspectives that disrupt stereotypical narratives. During decoding, CED locates the decoder layer exhibiting the greatest divergence among these perspectives and ensembles their token distributions using uncertainty-aware weights, prioritizing high-confidence tokens from different groups to yield a more balanced probability distribution that guides fairer generation. Extensive experiments on three social bias evaluation benchmarks demonstrate that \tool achieves substantial improvements over leading baselines, reducing bias by up to 47.97% across scenarios involving occupations, descriptors, and persona traits. Moreover, CED also preserves the core capabilities of the original model with minimal degradation.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Claude Desktop can now be configured to work with Ollama as a third-party gateway provider, making it possible to use open models in Claude.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Training and serving frontier AI models depends on fast, reliable networks that move data between GPUs without wasting compute cycles. To meet this challenge at scale, Meta designed MetaRoCE – a clean-sheet RDMA transpo…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Fairphone is finally selling its latest model in the US. The $649 handset has 11 swappable parts, and a replacement battery is only $40.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：You can now code your own mini games and 'gizmos' with no experience. Creating is just half the fun.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Metadata harmonization (standardizing labels, identifiers, and formats so datasets can work together) is still largely manual. This post shows how AI-powered metadata correction works in practice, covering two approaches, human-in-the-loop validation and autonomous agent-driven workflows, plus governance considerations for production deployment.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：The AI boom has disrupted the way engineers work, introducing new tools to learn, raising expectations for what teams can achieve in a workday, and making it harder to get hired in the first place. This makes it difficult to advise students on which specific coding languages or technical skills they should learn. So amidst the uncertainty, advice for young professionals often turns to a common refrain: Be adaptable. But what does adaptability look like in practice? Engineers often operate on the cutting edge of technology, so dealing with change is a normal part of the job, says Samantha Brunhaver, an associate professor of engineering at Arizona State University, in Tempe. Yet university curricula and training in the workplace often don’t prepare students for this. “We tell engineers that they need to be adaptable when they graduate, but we don’t actually explain what that means, demonstrate what that looks like, [or] help make sure that they’re developing it,” says Brunhaver, who received a National Science Foundation award in 2020 to study how to foster greater workplace adaptability among young engineers. For this ongoing project, she has interviewed engineering managers, early career employees, and undergraduates about their experiences. Part of the problem, she says, is that every employer has its own idea of what to be adaptable means. Generally, Brunhaver defines adaptability as “the ability to recognize that a change or uncertainty is occurring, and then respond effectively to that change.” But the skill is context-dependent. In software engineering, that might mean responding to turnover in the tools you use on a daily basis, while aerospace or biomedical engineers may need to keep track of changing procedures and regulations. “Managers are all saying adaptability is important,” Brunhaver says, “but defining it in different ways.” At the same time, engineers are all contending with changes beyond these industry-specific expectations. Jobs in the technology, media, and telecom sectors are experiencing the fastest pace of skill turnover, according to a June 2026 report on the effects of AI from the professional services network PwC. And the World Economic Forum’s most recent Future of Jobs Report, published in 2025, found that employers across all sectors expect 39 percent of workers’ core skills to change by 2030. This uncertainty can be uncomfortable. But with the right mind-set and support from leadership, adaptability can help keep you afloat. How to Cultivate Adaptability The AI transition is a big shift—but not an unprecedented one, says Jenna Butler, a research scientist at Microsoft who studies developer well-being and productivity. During this type of paradigm shift, there is often a “chaos period” when a new normal is being established, Butler says. In AI’s case, it challenges the understanding of what a computer can do. “I think we’re still in this in-between, difficult period that we’ve seen before, but [it] is maybe moving faster than it has historically.” Software engineers—in one of the fields most affected by AI—are now facing a significant increase in code review. “If you ask 20 developers, you get 23 different ways of working with it. Everyone is trying to sort it out,” says Butler, who describes this period as “the uncomfortable middle.” “We tell engineers that they need to be adaptable when they graduate, but we don’t actually explain what that means, demonstrate what that looks like, [or] help make sure that they’re developing it.”– Samantha Brunhaver, Arizona State University Brunhaver says one way educators can help prepare students before they enter the workforce is by offering a diversity of real-world experiences, such as internships, team-based projects, community service, and leadership roles. Each of these teach students to adapt to different challenges, easing their transition from school to work. It’s also important to encourage reflection, Brunhaver adds, noting that metacognition helps individuals use the skill more effectively. “In order to adapt, you have to think that you have agency and the ability to get through a situation.” Ultimately, it comes down to three steps: Perceive a need to adapt, evaluate your options, and act. For those already in the workforce, that action may mean taking the time to learn new tools and ways of working. Software engineering, for instance, may soon rely more on prompting models and managing agents than coding line by line. “I think people who went into software because they like solving problems are going to have a lot of fun, and people who just enjoy the art of writing code are not,” Butler says. The More Things Change… Although the tools engineers use on a daily basis are evolving, the core responsibilities of the job are more stable than they may seem, says Andy Hunt, a software developer who coauthored The Pragmatic Programmer (Addison-Wesley Professional) in 1999. The book outlines practical coding principles, and has been taught in many computer science classrooms. When Hunt was working on the 20th anniversary edition of the book, he was surprised by how much of the advice still applies. And now, seven years later, he maintains that belief. “The fundamental part of the job is problem solving and communication, and that’s always going to be there,” he says. Hunt emphasizes the importance of developing systems thinking over particular tools. To him, identifying as a Java programmer, for instance, is “like a carpenter saying, ‘I’m a hammer user,’ or ‘I specialize in cordless drills.’ ” He acknowledges that today’s hiring process, in which companies often filter résumés for certain languages or years of experience, makes it harder to embrace a more expansive way of relating to your job. Employers, he says, should recognize that “the tech’s not the hard part, and it never has been. Understanding information theory, understanding systems thinking, understanding what constraints you’re up to—that’s still the hard part.” With this type of misalignment between employers and employees, AI is also intensifying an old source of tension: How can engineers slow down enough to adapt and learn new tools when the pressure to become more productive keeps mounting? Who’s Responsible for Enabling Change? Young engineers need to embrace change. However, educators and employers also play a role in building a successful workforce. From the educator’s perspective, Brunhaver says “we need to be more explicit about what [adaptability] means and why it’s important.” Managers, meanwhile, should invest in their employees’ professional development. Microsoft research scientist Butler often encourages leadership to set aside intentional time for continuous learning for their engineers—even just an hour a week—without any expectation that they will produce code or progress in their daily work. “I realize that’s difficult,” says Butler. “I would encourage people to do it on their own, but I would really encourage organizations and leaders to do it, because you’re not going to get this sudden change in your people if they don’t have time and space to learn how to work differently.” This also means providing enough instruction, Butler adds. When developers aren’t given enough guidance on adopting something new, while being pressured to increase productivity, they risk doubling down on the tools they already know and burning out. “I do imagine the next number of years could be challenging,” Butler says. Engineers will have to adapt to find their place in an evolving workforce—but they also have a say in shaping that future. “Being adaptable sort of implies that you’re going to change based on what’s happening around you, and I would really like people to realize the change that’s happening is somewhat up to us,” she says. All individuals have a choice in how they use AI, for instance, and which models they use. “We need to be adaptable and go with the flow to a degree, but we also need to be directing that flow. The future with AI is absolutely not predetermined.” This article appears in the September 2026 print issue as “The Adaptable Engineer.”

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.20345v1 Announce Type: new Abstract: Conversational AI systems have become informal mental health support resources for Generation Alpha (Gen Alpha, born 2010-2024), with 13.1% of U.S. adolescents (5.4 million) using generative AI for mental health advice. While these systems, from therapy apps to general chatbots, rely on large language models trained on extensive psychological literature, their safety for youth communication patterns characterized by hyperbolic language, ironic positivity, rapid semantic drift, and contextual polysemy remains unvalidated. Following multiple adolescent deaths linked to AI chatbot interactions, systematic evaluation is critical. We present two benchmarks: (1) 64 Gen Alpha mental health expressions validated by native speakers (ICC=0.72) and clinicians (kappa=0.78); (2) 75 multi-turn conversations (780 turns) with paired Standard/Gen Alpha versions. Across evaluations of LLM architectures underlying therapy apps and general chatbots - Claude, GPT-4o, Llama-3.1 - models understand 76-82% of vocabulary but correctly calibrate only 64-72% of clinical risk, creating a 10-14 percentage point (pp) vocabulary-comprehension gap (p0.48) absent in human therapists (3pp, p=.22). The gap is architecturally consistent and widens with ambiguity (7pp -> 18pp). We identify six failure patterns: sarcasm masking (29pp), minimization acceptance (43pp), informal style bias (24pp), risk-stratified ambiguity (19pp), semantic drift (19pp), context-dependent violence (7pp). Patterns compound; three or more yield 94% miss rates. Lightweight mitigations fail; only heavy scaffolding achieves human performance (6.4x cost). With 34% baseline miss rate yielding 146,880 estimated annual missed crises, we recommend mandatory human-in-the-loop architectures, quarterly youth-specific validation, transparent performance disclosure, and regulatory frameworks for youth-facing mental health AI.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.20442v1 Announce Type: new Abstract: Subliminal trait transfer allows a student model to acquire behavioral dispositions from teacher-generated data in which the trait is not semantically expressed. Recent work explains how such signals enter gradients, but not how they survive source removal or acquire different signs under later training. We treat parameters and optimizer moments as a single trainer state and derive an exact transport-valuation identity separating observer-independent propagation of the source perturbation from the value assigned by a future continuation and behavioral readout. State surgery identifies the first moment as a causal carrier. Transplanting it alone leaves parameters, hidden states, and outputs unchanged at the cut, yet source-free updates generate growing parameter and hidden-state differences; transplanting parameters with the first moment recovers the terminal behavioral response. Sending the same source-induced difference through matched futures produces negative, near-zero, and positive Qwen effects (-0.658, +0.008, and +0.658 seed means). This ordering recurs in all 12 Llama-3.2-1B seeds after eight updates, while state-difference norms remain nearly equal across routes. Both contrasts grow in every paired seed when the continuation extends to sixteen updates. A full-horizon costate predicts all 42 Qwen route-mean signs and all 21 resolved Llama ordinary-route signs. Observer-independent transport also replicates across Qwen, SmolLM2, and Llama, while the complete-state recurrence predicts physical, hidden, and fixed-head responses in non-LoRA MNIST systems, including CNNs trained with AdamW and momentum SGD. Together, these results identify a two-stage mechanism for subliminal trait transfer: optimizer state transports the source perturbation, and later training determines its behavioral value.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.20343v1 Announce Type: new Abstract: This study develops and evaluates a bankruptcy prediction framework that integrates consensus-based feature selection, hybrid resampling, stacking ensembles, and explainable artificial intelligence to improve minority-class detection in severely imbalanced financial data. Using the Taiwanese Bankruptcy Prediction dataset from the UCI Machine Learning Repository, five feature-selection algorithms were first applied, and a consensus retention rule reduced the input space to 23 robust variables. The balanced training data were then generated using SVM-SMOTE, SMOTE-Tomek, and SMOTE-ENN. Five ensemble machine learning classifiers, namely gradient boosting, extreme gradient boosting, histogram-based gradient boosting, LightGBM, and AdaBoost, were compared with five deep learning models, including RNN, LSTM, GRU, DNN, and MLP. In addition, hybrid stacking ensembles combined the five machine learning classifiers as base learners with each deep learning model as a meta-learner. Model performance was assessed using accuracy, recall, specificity, G-mean, and ROC-AUC, while SHAP was used to explain feature contributions. The results show that resampling strategy materially shaped model behavior. SVM-SMOTE and SMOTE-Tomek favored accuracy and specificity, whereas SMOTE-ENN delivered stronger minority-class detection. Among standalone models, the GRU with SMOTE-ENN achieved the best overall predictive balance, with recall of 0.8627, G-mean of 0.8517, and ROC-AUC of 0.9431. Among stacking ensembles, SMOTE-ENN with (GB+XGB+HGB+LGBM+AB)+LSTM provided the strongest compromise between sensitivity and specificity. SHAP analysis identified leverage, profitability, solvency, and operational efficiency indicators as the most influential predictors of bankruptcy risk. These findings support more reliable and interpretable early warning systems for financially distressed firms.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.20341v1 Announce Type: new Abstract: Frontier coding agents backed by large language models with context windows from hundreds of thousands to millions of tokens are restructuring the Software Development Life Cycle (SDLC). Rich context handling and multi-step reasoning now allow substantial Functional Requirement Documents (FRDs) and repository context to be ingested in a single workflow, making specification quality the execution fuel for autonomous delivery. This report formalises Spec-Driven Agentic Development (SDAD) as a synthesis of disciplined up-front formalisation and high-velocity implementation: intent capture, machine-readable specification, agentic synthesis, and independent multi-agent verification under human sign-off. We revisit the historical pendulum between Waterfall and Agile, introduce AI-code as a fourth production paradigm, and compare Human-Agile (circa 2020) with Agentic-SDAD (circa 2026) across artefacts, cadence, accountability, and security posture. Beyond process description, we extend the model to team role metamorphosis (engineer, QA, platform, and product functions), quantitative governance (Ambiguity Tax, Spec Fidelity, SER, and TCI_agentic with repair multiplier phi), and pragmatic adoption via hybrid estimation and a staged migration blueprint. Industrial and research evidence on AI-augmented testing and verification is integrated to motivate separation between synthesis and release authority. Overall, the paper argues that agentic speed does not eliminate engineering discipline; it relocates discipline upstream into specification precision, explicit gates, and auditable provenance.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Fears of a datacenter buildout debt crisis are exaggerated. The risks are different than in the past and they are recoverable Some experts are warning of a looming “debt bomb” crisis because big datacenter builders such as Meta, Oracle, xAI and CoreWeave are not only raising billions to construct these facilities but are also not recognizing these long-term debt obligations on their balance sheets. Should we be concerned? No. I know because I’ve seen this movie before. Continue reading...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Chris Lehane tells Guardian of need to implement new safety standards as critics say AI firms acting ‘recklessly’ A senior leader at OpenAI has said people should prepare to defend against “ongoing, persistent” cyber-attacks from AIs, as cutting-edge artificial intelligence models gain advanced capabilities to plan and launch offensives. The leading AI company this week announced a pause in development of its most advanced internal models amid rising safety fears, and Chris Lehane, its chief global affairs officer, said: “We are hitting a different chapter, a different moment within AI, in terms of what the capabilities of this technology can do.” Continue reading...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Between 7-16 August, an AI agent (Claude Opus 5) has been acting as a desk officer at the Ministry for Foreign Affairs of a fictional state called Sordland (from Suzerain, not sponsored, but buy it, its a good game). It…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：“I can do something about it” Since February, engineers have been releasing and updating apps that help people detect glasses like Ray-Ban Meta, Oakley Meta, or Snap Spectacles. “It is completely unsurprising that peopl…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：What's up, ChickenButt? I made a free, native GTK client called ChickenButt. :) It lets you chat with local models through Ollama, and I figured some of you might get a kick out of it. Here's the repo: https://github.co…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Run Muse Glimmer locally on an RTX 3090 GPU using llama.cpp, DFlash speculative decoding, and Pi for fast, private, agentic AI coding.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：siriusly.ai | UGC maxxing UGC maxxing. Multiply your content. Turn a product image or screen recording into dozens of viral UGC reels, optimized for Meta Andromeda. No face-swap, no lip-sync - natively shot per language…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.19201v1 Announce Type: new Abstract: Bioinformatics software and databases are essential components of modern life science research, yet their mentions in the scientific literature are often inconsistent and difficult to systematically identify at scale. The lack of a comprehensive and up-to-date catalog of bioinformatics resources hinders efforts toward automated biomedical knowledge extraction and streamlined data analysis. Here we present SNAIL, a hybrid named entity recognition framework designed to automatically identify bioinformatics software and database (SW/DB) names from biomedical texts. SNAIL integrates complementary lexical and semantic modeling strategies. The lexical component captures orthographic patterns and contextual cues characteristic of SW/DB names, while the semantic component leverages contextual embeddings generated by transformer-based language models such as SciBERT, combined with an explicit token-masking strategy to enhance entity-focused representations. A large training corpus was constructed automatically through a hybrid pipeline that integrates citation-hinted extraction with large language model-assisted distillation. Evaluation on two independent benchmark datasets and real-world research articles demonstrates that SNAIL substantially outperforms existing approaches, including domain-specific methods such as bioNerDS2 and general-purpose large language models such as ChatGPT, Gemini, Grok and Claude. Applying SNAIL to large-scale literature analysis further reveals distinct journal-level preferences across bioinformatics subfields. These results demonstrate that SNAIL provides an accurate and scalable solution for identifying bioinformatics resources in scientific texts and enables systematic meta-analysis of tool usage and research trends.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.19447v1 Announce Type: new Abstract: Shocks that spread through the web, such as cybersecurity breach disclosures, can abruptly disrupt financial time series and cause substantial abnormal losses. While these events are disclosed as discrete records through news reports, regulatory filings, or public databases, their consequences unfold through continuous market dynamics. This creates an event-conditioned impact prediction problem: given pre-event market history and limited event metadata, the goal is to estimate short-term post-disclosure abnormal loss rather than reconstruct the full post-event trajectory. However, most time-series forecasting models focus on endogenous regularities such as trend, seasonality, and autocorrelation, and thus struggle with rare and heterogeneous external events. The challenge is further amplified by sparse high-impact events and background market noise. We introduce EventTime, a multi-resolution framework that combines long-horizon market context, short-horizon pre-event dynamics, and event metadata. It incorporates an event fusion module that couples temporal representations with event attributes to identify relevant recent market patterns. To mitigate sparse supervision, EventTime further introduces a dynamic contrastive objective that constructs event- and time-series-aware positive and negative pairs during training. We also construct SECURE, a real-world dataset aligning cybersecurity incidents with stock-market time series and structured and LLM-derived semantic features. Experiments show that EventTime consistently outperforms state-of-the-art time-series and event-aware baselines in estimating post-event financial losses. Further analyses demonstrate more event-sensitive representations, greater robustness to incomplete metadata, and more interpretable estimates of short-term market impact following cybersecurity disclosures.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.19535v1 Announce Type: new Abstract: Retrieval-augmented generation (RAG) improves language-model responses by grounding generation in external passages, which comes with overhead: retrieved context lengthens the prompt, increasing prefill work, KV-cache footprint, memory traffic, latency, and energy. Context compression offers a natural remedy by pruning retrieved text before generation. However, state-of-the-art context-compression methods are typically used with a fixed compression budget, or with the rate selected offline and then applied at inference time. This static view ignores both workload variation and the live state of the edge device. On an edge SoC, compression is not free: the compressor itself runs on the same SoC and consumes latency and energy that can offset any generation savings. This paper proposes a vision for telemetry-informed adaptive compression in edge RAG, grounded in experimental evidence. We characterize the compression tradeoff on the NVIDIA Jetson AGX Thor using Llama and Qwen generators, Natural Questions and HotpotQA datasets, and LLMLingua-2 compression. Our measurements show that generation dominates the RAG budget for larger models, reaching roughly 90% of per-query latency and 91% of GPU energy for 7B-8B generators. Exploring the impact of the compression rate reveals an adaptive operating region: mild compression can miss energy opportunities, and overly aggressive compression can hurt inference quality. Intermediate compression can reduce GPU energy by up to 53.2%, and SoC energy by up to 48.2%, with negligible quality loss. We argue for runtime policies that dynamically manage compression, guided by workload features and edge telemetry.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.19214v1 Announce Type: new Abstract: The berth allocation and quay crane assignment problem (BACAP) is a representative port-terminal scheduling problem in maritime transportation and freight logistics, where vessel arrivals, berth positions, service durations, and quay?crane availability are tightly coupled. Under uncertainties such as arrival deviations, handling-time fluctuations, and resource disruptions, schedules optimized under nominal assumptions may become fragile during execution, motivating the study of robust metaheuristic optimization for BACAP in port-terminal operations. Although population-based metaheuristics have been widely used for BACAP and related port-scheduling problems, existing studies remain fragmented in their uncertainty repre?sentations, robustness criteria, search mechanisms, and empir?ical evaluation protocols. To the best of our knowledge, this paper provides the first focused review dedicated to robust population-based metaheuristics for BACAP under uncertainty. We first summarize uncertainty sources and information repre?sentations in BACAP, and then organize existing methods from a mechanism-oriented perspective, covering solution representation and decoding, robust evaluation and selection, robustness-guided search dynamics, and feasibility preservation and recovery. We further present a benchmark suite for uncertain BACAP to support controlled empirical comparison and report illustrative baseline results by combining representative metaheuristics with different robustness strategies. Finally, we identify open chal?lenges related to benchmark extension, robustness-aware search design, time-adaptive robustness, and non-stationary uncertainty.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Discussion | Link

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：The release is part of the company’s strengthened emphasis on AI and appealing to businesses.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Trade body says local chains would have to balance concerns with potential benefits provided by AI-enabled technology Cinemas across the UK are considering banning customers from wearing Meta’s smart glasses, amid fears they could be used to pirate films. The UK Cinema Association (UKCA) trade body said that a number of local chains could end up introducing policies that restrict camera-enabled smart glasses at the big screen, adding to a growing list of venues concerned about covert use. Continue reading...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：--> [Submitted on 27 Mar 2026 (v1), last revised 6 Apr 2026 (this version, v2)] Title:Beyond BMI: Smartphone Body Composition Phenotyping for Cardiometabolic Risk Assessment View a PDF of the paper titled Beyond BMI: Sm…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18309v1 Announce Type: new Abstract: Registering images acquired with different microscopy modalities is essential for relating complementary measurements of the same specimen. In correlative X-ray fluorescence (XRF) and optical microscopy, the XRF map often covers only a small region of an optical image acquired from the same or an adjacent tissue section. Field-of-view (FOV) localization is necessary but can be difficult when appearance and structure differ across modalities. Here we evaluate training-free vision language model (VLM) localization on two datasets representing same-section high-correspondence and adjacent-section low-correspondence imaging. We test unconstrained and metadata-constrained search and compare VLMs with geometric controls, classical template matching, and two alternative training-free approaches (DINOv2 and multiGradICON). Direct VLM prompting produced content-dependent spatial signals but was not reliable alone. Classical matching was most accurate when cross-modal structure was preserved but failed in the low-correspondence collection. A proposal-and-verify workflow used repeated VLM predictions as candidates and image-based similarity to select the final location. This workflow recovered useful localization in the low-correspondence regime.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18305v1 Announce Type: new Abstract: X-ray fluorescence (XRF) microscopy maps elemental distributions, while optical microscopy can provide complementary morphological context. Localizing XRF fields of view (FOVs) in optical images is difficult because the two modalities differ in contrast mechanism and resolution. Most current workflows place each XRF tile independently, even when acquisition metadata already record the tiles' relative scan positions. This study formalizes XRF tile-group localization, in which one optical-frame placement is estimated for the whole group, constrained by acquisition geometry and quantified using group intersection-over-union (GroupIoU). In a controlled case study, independent localization failed with GroupIoU 0.000, whereas group localization achieved 0.931. Replacing the normalized cross-correlation (NCC) metric with mutual information (MI) gave nearly identical results, showing that the outcome is not specific to one local similarity metric. In another multiscale case study, using a coarse XRF survey scan to connect the fine-scale tile group to the optical image increased mean GroupIoU from 0.694 to 0.856. These case studies support using acquisition geometry as an explicit constraint when localizing related XRF tiles.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18093v1 Announce Type: new Abstract: Abliteration, the removal of refusal capabilities from large language models by projecting weight matrices orthogonal to an extracted refusal direction, has emerged as a prominent safety concern through its ability to bypass post-training alignment using only a small set of contrastive prompts. We find that existing defenses commonly overlook the cause of abliteration; that is, how easily the refusal direction can be extracted. To hinder this process, we introduce a weight-editing method that obscures the refusal signal by applying rank-$k$ updates to residual stream writer matrices while replacing refusal-inducing activations with random aliases and correcting downstream reader matrices to preserve the model's original behavior. On Llama-3-8B, AMRA improves post-abliteration refusal scores by $2.16$ points over the undefended baseline with less than $0.5$ percentage points of MMLU degradation. On Gemma-2-9B, it improves the post-abliteration refusal by $14.70$ points over the baseline while keeping harmful output rates similar to the baseline, albeit at a greater utility cost.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18090v1 Announce Type: new Abstract: Inside a modern language model sits a single internal direction that tracks how positive or negative a sentence feels. We show how to find this valence axis (V-axis) from just 9 emotion category names plus 50 short narrative paragraphs per emotion -- about 1,500 fewer labels than the usual supervised approach -- and that the same direction appears in vision, audio, and human-brain encoders never jointly trained. The recipe: embed nine emotion-anchored story sets in a frozen encoder, take the top principal direction of the nine averaged embeddings. Projecting new inputs onto it captures 93% of supervised performance on SST-2 (Llama-3-8B-Instruct, AUC 0.772 vs. 0.828), correlates with human valence ratings on 11,811 EmoSet images at r=0.636, reaches AUC 0.906 on ESC-50 audio (p12). A 2-parameter classifier trained on text labels transfers to images (AUC 0.961), audio (0.764), and brain recordings (0.828) without target-modality labels; a generic 16-D subspace stays at chance (0.525). The recipe is bounded to continuous attributes -- seven tests on categorical concepts return near-chance -- and steering is family-specific (Llama/Mistral yes, Qwen/Gemma no).

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18089v1 Announce Type: new Abstract: Instruction-tuned models often refuse harmful requests in English but comply with the same requests in Yoruba, Igbo, Igala, and Hausa. This suggests that the refusal mechanism is present in the residual stream but fails to activate for low-resource inputs. Recovering it normally requires labelled target-language data and retraining, neither of which is available at scale for most African languages. We introduce Latent Space Refusal Anchoring (LSR-Anchoring), a training-free method that extracts the refusal direction from English prompts and clamps it onto the residual stream at inference time. The primary variant, Mean-Activation Steering (MAS), operates across the four architectures we tested: Llama-3-8B, Llama-3.1-70B, Mistral-7B-Instruct, and Qwen2.5-7B. On Mistral and Qwen it recovers safety with benign degradation below 0.08. On Llama-3-8B it overcorrects, with Degraded Performance on Legitimate prompts (DPL) reaching 1.00. We address this with SAE-Derived Steering (SDS), which replaces the dense mean-difference direction with a single Sparse Autoencoder (SAE) feature and reduces Kullback-Leibler (KL) divergence by 3.5-7x without benign collapse. Four languages transfer positively, but Arabic fails on every architecture and at every steering magnitude, indicating a geometric mismatch rather than a baseline effect. Massive Multitask Language Understanding (MMLU) accuracy drops remain below 0.35 percentage points at every effective steering magnitude.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18183v1 Announce Type: new Abstract: Visual on-policy distillation (OPD) improves the training of compact visual autoregressive models by learning from trajectories generated by the current student. However, these online rollouts are still produced token by token with autoregressive decoding, which adds substantial cost to every on-policy training step. Speculative Jacobi Decoding (SJD) provides an alternative because it can process multiple tokens in parallel without an auxiliary draft model, but the original method is designed for single-sequence inference. We introduce HB-SJD, a batched SJD rollout backend for visual OPD. HB-SJD allows each image to advance independently according to its own decoding progress, while images at different sequence positions are still verified in batched model forwards. As images finish, HB-SJD switches between Full and Compact execution to reduce the cost of later rollout rounds. HB-SJD only replaces the student rollout backend and leaves the teacher, distillation objective, and optimization procedure unchanged. Experiments with LlamaGen show that HB-SJD substantially reduces rollout and end-to-end training time while preserving the generation quality of the distilled student.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18088v1 Announce Type: new Abstract: Drone propeller faults can create safety and reliability risks when their effects are distributed across multiple flight-log channels rather than appearing as a single diagnostic signal. This paper proposes a Metamorphic Artificial Age Score (AAS) decision-support prototype for flight-log-based drone propeller health monitoring. Using selected historical real flight logs from the 2024 DronePropA public dataset, the framework computes six health-related indicators from raw MATLAB matrices: trajectory tracking error, attitude instability, thrust-command burden, motor-command imbalance, ESC-command instability, and battery-level stress. These indicators are normalized relative to a healthy baseline and evaluated through candidate scoring policies, metamorphic adequacy relations, and a redundancy-adjusted AAS formulation. In this context, AAS is used as a structural policy-adequacy and burden measure rather than as a chronological age measure. A controlled retrospective evaluation was performed using one healthy baseline and three defective propeller cases under the same speed profile and trajectory. The healthy case was assigned to routine monitoring. The Severity 1 case was dominated by ESC-command instability and assigned to maintenance review. The Severity 2 case reached maximum motor-command and ESC-command burden, while the Severity 3 case reached maximum trajectory tracking error; both triggered mandatory inspection. The results show that propeller fault effects may appear through different operational channels, supporting the need for a multi-indicator decision-support layer for post-flight maintenance prioritization and autonomous-system oversight.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.18079v1 Announce Type: new Abstract: Game world modeling (GWM) and reinforcement learning (RL) are often confounded because research papers rarely quantify how difficult the underlying transition prediction problem is at the declared interface (pixels/tokens/latents with finite history). We propose the Transition Complexity Profile (TCP): a small, reproducible set of metrics that characterizes an environment's (or gameplay dataset's) induced transition kernel by (i) intrinsic one-step branching, (ii) interaction-induced uncertainty and opponent influence when observable, and (iii) temporal/spatial dependency span via standardized probe curves. TCP is reported with an explicit reference distribution, protocol stochasticity, and a versioned measurement budget (sampling/resampling and fixed probe compute), enabling comparable numbers across benchmarks. We outline how common game families and modern "neural game engine" domains populate this landscape and call for TCP to become standard benchmark metadata and a required statistic in GWM and RL papers.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：8/18/2026 Mark Zuckerberg's AI Manifesto Shows How Little The AI Future Offers Mark Zuckerberg of Meta fame posted a manifesto. He lays out an idea of a "future for everyone", although I don't think his group of everyon…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：An Apple spokesperson told WIRED that the company has removed Kromix—which, they said, appeared to have added prohibited content and features after initial review—from the App Store. “We have always strictly prohibited…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Meta AI can create content and make suggestions based on what it “sees” on your screen. | Image: Meta Meta is launching a new Mac app dedicated to its AI chatbot. In an announcement on Wednesday, Meta says you can share your window with its AI chatbot, which can provide suggestions, answer questions, or create content based on what's on your screen. Meta AI on the Mac also supports dictation across all apps. The launch comes as Meta makes its AI chatbot more like a productivity-focused assistant to better compete with its AI rivals - many of which already have desktop apps. You can share your window with Google's Gemini AI app, while OpenAI's ChatGPT and Anthropic's Claude apps take things a step further by allowing the chatbots to take con … Read the full story at The Verge.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Fairphone is finally selling its latest model in the US. The $649 handset has 12 swappable parts, and a replacement battery is only $40. This is truly the Goldilocks of Android phones.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：People say they’ve been secretly filmed in their own home, at concerts and at work. Are the wildly popular smartglasses the final nail in the coffin of personal privacy? “I’ve had one person who told me that their intentions were creepy,” a man tells me over a video call, on condition of anonymity. He’s based in Los Angeles, and while we speak, he eats what appears to be tuna directly out of the can. “He said: ‘I go to strip clubs, and I want to record the strippers … Normally I put my phone in my chest pocket, but the glasses are more convenient.’” The man I’m talking to runs a business called Ghost Metas. He’s one of hundreds of vendors, easily discoverable online, who specialise in disabling the flashing LED light embedded in Meta’s smartglasses that blinks when wearers capture photos, videos and audio. After Ghost Metas disables the LED, it’s impossible for someone to know they’re being filmed. Continue reading...

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Claude now marks AI-generated content. But it does not mark everything the same way. Anthropic currently uses embedded watermarks for text and signed C2PA provenance metadata for supported files. Code sits somewhere in between: it is still text, but its structure gives the watermark fewer places to work. I went into detail about Claude’s watermarks […] The post How to Remove Claude Watermarks from Text, Code, and Files appeared first on Analytics Vidhya.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.16973v1 Announce Type: new Abstract: Blueberry ripeness is judged by berry colour, cluster composition, and the distribution of maturity stages within a plant, however, public green house image resources with dense ripeness-stage masks remain limited. We present AerialYield-B2D, where B2D denotes BlueBerry Dataset, acurated real-image resource containing 514 RGB images and 30,195 annotated blueberry instances across five ripeness stages: green immature, pale pink, pink-turns-purple, fully ripe and over-ripe. The release provides class-specific binary masks, overall berry masks, semantic label maps, image-level count tables, SHA-256 hashes, source metadata, recommended train/validation/test splits and technical validations. AerialYield is the broader project name; this release does not provide harvest weight, fruit mass or per-area yield measurements, and the count labels should therefore be interpreted as image-level berry counts rather than yield estimates. The images include 424 smartphone greenhouse images, 67 video-derived frames, and 23 DJI Fly video-frame samples, providing a reproducible dataset for ripeness segmentation, berry counting, and class-imbalance analysis in controlled-environment blueberry production.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：arXiv:2608.16956v1 Announce Type: new Abstract: API buyers purchase a dated contract, not a model name alone: the contract includes the requested and served model, reasoning-effort term or its omission, output rail, service product, prompt, and price schedule. We study the reasoning-effort term through a registered paired contrast of Sonnet 5 with explicit high effort against the same model with effort omitted, using 30 AIME 2026 items and five calls per item. Every paid attempt was assigned one frozen terminal category, and inference resampled items while retaining their repeated calls. Mean delivered cost was \$0.01031 per call higher under the explicit-high contract than under the omitted contract [+\$0.00204, +\$0.01974]. The corresponding accuracy contrast was +0.0133 [-0.0267, +0.0467]; we did not detect an accuracy difference, and the interval permits a gain of up to 4.67 percentage points that this design cannot rule out. Cost per correct answer was \$0.08665 under the high-effort contract and \$0.07662 under the omitted contract, as registered point estimates. A dated contract census, Models-API metadata, and preregistered raw-response probes further documented model-specific omission semantics, including within a provider; claims remained at documentation grade when raw structure was indeterminate. The request registry, parser, terminal taxonomy, statistical plan, and analysis pipeline were frozen before outcomes were examined; the resulting claims are bounded to the model, task, and collection date studied.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Codex alternative for Open Source AI curl -fsSL https://chatoss.ai/install.sh | sh paste this in terminal, or download ChatOSS Built for agentic coding ChatOSS is a native desktop workspace where AI agents do real work…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：In this post, we describe how AIDA works at a high level and how it helps address these challenges — grounding users in the right contracts, under the right legal context, and within the right access boundaries. Specifically, we explore how AIDA uses implicit and explicit filtering, along with metadata-enriched chunking in Amazon Bedrock Knowledge Bases, to dramatically improve contract search accuracy.

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：The rotation spreadsheet always rots The pinned spreadsheet nobody updates Last edited eleven months ago. Two of the names don’t work here anymore. The /remind hack that pings people on vacation Nothing says “fair” like…

AI 服務暫時不可用，以下為來源摘要，待恢復後補全翻譯：Download Ollama, pull and serve Qwen3.8-27B, and launch it with OpenCode using just three command lines.