arXiv:2607.01282v1 Announce Type: new Abstract: In light of strides in Arti cial Intelligence (AI) and its wide spread application, challenges persist in the interpretability of AI models, particularly within specialized domains like healthcare, such as electro cardiograph (ECG) recognition. Rather than relying solely on end-to-end convolutional neural networks, this paper introduces a novel approach using a domain knowledge-based graph convolution network for ECG recognition. Key landmarks points of PRQST, vital to ECG interpreta tion, are incorporated as domain knowledge. The double-stream directed graph is employed to model both intra and inter ECG cycles. Speci cally, spatial directed graphs capture the positional relationships among key points, while temporal directed graphs delineate temporal dependencies between adjacent cycles in extended ECG sequences. Experimental re sults on the First Chinese ECG Intelligent Competition dataset, which speci cally classify ECG into nine categories, prove the e cacy of the proposed model. The overall average F1 score is 88.1%, the average F1 score of rare categories is 76.3%, both outperform the state-of-the-art models. The introduction of domain knowledge did enhance the detec tion performance, especially for rare categories.

arXiv:2607.01283v1 Announce Type: new Abstract: Grid-based approaches to approximate nearest neighbor (ANN) search have been absent from modern scaling analyses. We present a systematic characterization of a multiprobe grid algorithm with respect to dataset size $N$ and dimensionality $d$. Our experiments reveal a previously unreported $d$-scaling crossover on the GloVe embedding family, in which multiprobe grid search maintains an approximately constant dimensional scaling exponent while other graph-, tree-, and partitioning-based methods exhibit degrading throughput. The advantage comes with near-linear query scaling in $N$, but also with lower indexing cost than competing ANN methods. Our results suggest that grid-based methods such as multiprobe grid may be competitive in rebuild-heavy or high-dimensional settings where indexing cost and dimensional robustness dictate performance. More broadly, recent work has formalized self-attention as an ANN operation. Thus, the $N$- and $d$-scaling properties of ANN algorithms may guide cost analysis of efficient transformer architectures. Code is available at: https://github.com/weiz345/MultiProbeANN.

arXiv:2607.01286v1 Announce Type: new Abstract: Public lithium-ion battery datasets are increasingly used for state-of-health estimation, remaining-useful-life prediction, anomaly detection, electrochemical diagnostics, second-life analytics, and battery safety research. However, these datasets vary substantially in chemistry, modality, scale, label quality, sequence structure, access status, and preprocessing complexity. These differences directly affect whether a dataset is feasible for near-term hybrid quantum-classical machine-learning workflows. This paper presents IonSense-QKG, a quantum-readiness metadata framework for lithium-ion battery dataset discovery. Starting from the EV-Battery-IonSense index, the proposed framework enriches public battery dataset records with quantum-relevant metadata, including task type, sensing modality, chemistry, label availability, sequence type, preprocessing requirements, candidate quantum encodings, estimated qubit range, and NISQ feasibility. A transparent Quantum Readiness Score is introduced to rank datasets as candidate resources for future hybrid quantum-classical battery benchmarks. The score is intended as a dataset-selection heuristic, not as evidence of quantum advantage. The framework demonstrates query-based discovery over enriched metadata to identify datasets suitable for compact quantum feature maps, quantum time-series workflows, limited-label anomaly detection, and future battery-health benchmarking. The released artifact includes metadata tables, scoring scripts, robustness checks, link-checking utilities, and SQL-style query examples. IonSense-QKG positions dataset selection as a data-management problem and provides a reproducible foundation for data-centric quantum battery analytics.

arXiv:2607.01311v1 Announce Type: new Abstract: Deep learning has outgrown any single mathematical explanation. From Approximation to Emergence develops a unified, proof-oriented account of modern deep learning theory, tracing a path from the classical foundations of approximation, optimization, and generalization to the contemporary mechanisms of overparameterization, robustness, generative modeling, transformers, in-context learning, scaling laws, interpretability, alignment, and emergence. Rather than presenting isolated results, the book organizes a broad literature into a coherent research narrative: each theory is examined through the object it controls, the assumptions that make it valid, and the phenomena it leaves unexplained. Written for researchers, graduate students, and mathematically trained practitioners, this monograph offers a rigorous map of deep learning theory as it stands today: powerful, incomplete, and increasingly centered on the question of how learned mechanisms arise from scale, data, architecture, and training.

arXiv:2607.01235v1 Announce Type: new Abstract: Understanding how Large Language Models (LLMs) make token-level decisions during code generation remains a major challenge for both researchers and practitioners. While recent tools provide insights into model internals or generation outcomes, they often lack decoding-time signals, fine-grained uncertainty measures, and interactive mechanisms for exploring alternative generation paths. We present TokenScope, an interactive interpretability and analysis tool for decoder-based LLMs that exposes token-level metrics, attention patterns, and structural information during generation. TokenScope supports interactive token replacement, counterfactual branching, and code-aware aggregation via abstract syntax trees. By unifying decoding-time signals with structural program analysis, TokenScope enables systematic investigation of LLM behaviour during code generation.

arXiv:2607.01236v1 Announce Type: new Abstract: As LLM agents gain increasing access to powerful tools, ensuring that their actions are aligned with the user's intent becomes critical. When an agent's proposed tool invocation deviates from the user's intent -- a phenomenon called misalignment -- it may lead to harmful consequences that are difficult to undo. Existing runtime guardrails rely on an LLM-as-a-judge paradigm that lacks a systematic framework for reasoning about alignment, often producing judgments that are inconsistent or difficult to audit. Motivated by provenance analysis, we propose a provenance-based conceptual framework that formalizes misalignment detection as determining whether a proposed tool call is supported by traceable evidence in the agent's context. Building on this framework, we propose ProvenanceGuard, a multi-stage pipeline that analyzes the agent's action for three types of misalignment before the selected tool is executed and only allows the action to take place when it is considered aligned with the user's input query. We evaluated our proposed approach on two different benchmarks, Agent-SafetyBench and WorkBench, across 10 backbone LLMs. Compared to the LLM-as-a-judge baseline, ProvenanceGuard reduces error rate on misaligned traces from 42.9% to 1.8% on Agent-SafetyBench and from 32.1% to 17.3% on WorkBench, while reducing intervention burden on task-successful traces from 30.5% to 12.8% and introducing no statistically significant increase in unnecessary interventions on aligned traces. These results demonstrate that structured, provenance-based reasoning provides an effective and practical foundation for safeguarding LLM agents from misalignment.

arXiv:2607.01237v1 Announce Type: new Abstract: Reasoning language models often generate long chain-of-thought (CoT), which accumulates a massive KV cache during the decoding phase and incurs high decoding latency and limited throughput. To address these issues, KV cache compression has emerged as a promising technique for reducing memory overhead by selectively removing unimportant KV pairs while preserving useful ones for subsequent decoding. Nevertheless, we identify two key limitations in existing KV cache compression methods: 1) their threshold-triggered compression policy may provide limited throughput improvement or even reduce throughput, and may fully eliminate KV pairs from certain blocks of the sequence, potentially worsening information loss. 2) they typically retain either isolated KV pairs or fixed-size chunks with rigid boundaries, failing to preserve important flexible-sized chunks at arbitrary token positions. To overcome these limitations, we propose Kara, a sliding-window KV cache compression method that performs decoding-time compression by operating only on the recently generated context. Kara leverages bidirectional attention to score and select informative KV pairs in the window. To enable flexible preservation of important semantic information, we design a Token2Chunk module to expand a subset of selected KV pairs into chunks. Furthermore, we adapt Kara to PagedAttention and develop KvLLM, an inference framework built upon vLLM, which reduces KV cache memory usage and effectively improves output throughput. Extensive experiments demonstrate consistent performance improvements of proposed Kara and KvLLM.

arXiv:2607.01240v1 Announce Type: new Abstract: Count-based F1 is widely used as a proxy for LLM error-detection quality, but this paper shows that it can rise dramatically without a corresponding improvement in span localization, a gap termed F1 Inflation. The paper introduces ErrorBench, a controlled stress-test protocol for prompt-induced count distortion. ErrorBench evaluates six contemporary LLMs under five prompt conditions over 4,290 responses from 143 CoNLL-2014 passages. Under CoNLL-2014 M2-style scoring, anchored prompts produce up to 0.79 points of F1 Inflation, and up to 0.96 under strict matching. A 100-passage replication using the official ERRANT 3.0.0 pipeline and multi-reference scoring reproduces the pattern: averaged over six models, the Blind-to-Anchored prompt shift raises Count-F1 by +0.21 while raising multi-reference ERRANT F0.5 by only +0.04. The study finds larger count responses in highly instruction-compliant GPT/Claude systems and smaller responses in the Gemini family under this stress-test protocol. The findings suggest that LLM proofreading and document-review evaluations should avoid pre-populated error counts and should report span-aware metrics alongside count-based metrics.

arXiv:2607.01241v1 Announce Type: new Abstract: Existing prompt compression methods treat text as flat token sequences, failing to capture the distributed nature of important information, which is often spread across multiple locations and connected through both local syntactic dependencies and global semantic relations. Such relational structure is naturally represented as a graph, where tokens or sentences become nodes and their dependencies become edges. To this end, we propose RAGP, which formulates prompt compression as Redundancy-Aware Graph Pruning on a multiplex graph that jointly models fine-grained attention-based dependencies and coarse-grained semantic relations. To efficiently identify non-redundant nodes in this heterogeneous structure (dense local subgraphs and sparse global connections), we employ Levy walks whose heavy-tailed step distribution naturally balances local exploitation with global exploration. Experiments on LongBench show that RAGP achieves an average score of 49.3 under a 4x compression ratio, outperforming existing LLM-based compression methods, such as LongLLMLingua, which attains 48.8 at a 3x compression ratio. Besides, RAGP also surpasses state-of-the-art vision-based text compression paradigms on multiple tasks. The code is available at https://anonymous.4open.science/r/RAGP-B0CB.

arXiv:2607.01245v1 Announce Type: new Abstract: We introduce Office Comprehension Bench (OCB), the first public benchmark to jointly evaluate LLM systems on Word, Excel, and PowerPoint comprehension over native file formats (.docx, .xlsx, .pptx) and their variants. OCB consists of two tracks. File Fidelity Q&A tests structural and visual perception of office artifacts - tables, charts, embedded images, formulas, and app-specific elements such as headers, speaker notes, and named ranges. Domain Q&A tests expert-level reasoning grounded in real-world industry documents across 12 professional domains, with queries requiring multi-step analysis and synthesis across documents. Each reference answer is decomposed into atomic, binary-gradable claims, and an ensemble of LLM judges scores responses against each claim independently. Even the strongest frontier system in its default reasoning mode reaches only about 59.3% on Domain Q&A; increasing thinking depth within a tier does not move performance materially, while moving to a higher product tier yields modest gains. We release the dataset, evaluation tooling, judge prompt, and a public leaderboard.

arXiv:2607.01290v1 Announce Type: new Abstract: 3D Gaussian Splatting (3DGS) has emerged as a powerful representation for high-fidelity rendering. However, existing assets often suffer from quality bottlenecks such as missing details and texture noise. Prior attempts to enhance these assets via 2D image processing introduce multi-view inconsistencies and high computational costs. In this paper, we propose a novel 3D-native refinement paradigm named AnchorSplat. AnchorSplat is an end-to-end deep network operating directly on 3D structures, avoiding the expensive optimization overhead of traditional 3D-2D-3D pipelines. Crucially, AnchorSplat is a strictly source-free solution requiring no original multi-view images. Central to the proposed method is the Point Anchor Mechanism, which enforces geometric consistency via local offset constraints, mitigating ill-posed mapping and gradient confounding. Furthermore, AnchorSplat replaces iterative densification with a single-pass multiplication mechanism. To facilitate research, we construct 3DGS-SR, the first large-scale benchmark for this task. Experiments demonstrate state-of-the-art results on the 3DGS-SR dataset, with throughput up to $10^5$ times faster than optimization methods. Notably, AnchorSplat exhibits robust zero-shot generalization across diverse data distributions, including generative model outputs and real-world scans.

arXiv:2607.01303v1 Announce Type: new Abstract: Presentation Attack Detection (PAD) serves as a crucial safeguard for face recognition systems against presentation attacks such as printed photos, replayed videos, and 3D masks. Despite significant progress, existing PAD models still struggle to generalize across unseen domains due to variations in sensors, lighting, and attack materials. Recent Vision-Language Models (VLMs) have shown strong generalization ability, yet their applications in PAD remain limited because learned prompts, typically optimized under class-label supervision, fail to explicitly align with fine-grained attack-relevant visual semantics. As a result, the learned representations often overfit domain-specific artifacts instead of capturing transferable attack cues. To address this, we propose Concept-Informed Prompts Guided Presentation Attack Detection (CPG-PAD), a framework that introduces model-level concept guidance into the prompt learning process. Specifically, we design a Visual Concept-driven Enhancement (VCE) module that employs eXplainable AI (XAI) techniques to automatically discover PAD-relevant visual concepts and generate concept-associated heatmaps providing localized fine-grained guidance. Guided by these heatmaps, a Prompt-based Concept Injection (PCI) mechanism integrates these concepts into the prompt space through a Visual-Prompt Decoder (VPD) and a concept-mapping loss, enabling prompts to align with the model's internal concept space. This design enables CPG-PAD to capture generalizable and domain-invariant attack cues while effectively suppressing dataset-specific biases. Extensive experiments across nine benchmark datasets demonstrate that CPG-PAD consistently achieves state-of-the-art cross-domain performance under multi-source, limited-source, and single-source settings.

arXiv:2607.01278v1 Announce Type: new Abstract: The research proposes a multilayer Q-matrix-embedded neural network for cognitive diagnosis (M-QCDNet), which integrates the structural interpretability of cognitive diagnostic models (CDMs) with the deep learning neural network (NN). M-QCDNet structures the item-skill relationship using the Q-matrix as a structural prior, ensuring latent mastery profiles remain interpretable and consistent with cognitive theory, followed by the proposed loss function with an L2 penalty to penalize skills not aligned with the Q-matrix and to balance predictive performance and structural alignment. Corresponding evaluation matrices, the interpretable alignment-based metrics that quantify the degree to which predicted skill activations correspond to item-level skills, were further developed. M-QCDNet offers practical benefits for classroom practice, enabling early detection of learning difficulties and supporting mastery-based interventions. By embedding diagnostic validity into model design, M-QCDNet bridges psychometric transparency and neural flexibility, advancing interpretable, fair, and actionable AI for cognitive diagnostics.

Travis Kelce and Taylor Swift attend the 2026 iHeartRadio Music Awards. Kevin Mazur/Getty Images for iHeartRadio Taylor Swift and Travis Kelce donate $26 million to charities before their NYC wedding. The couple's donations support charities in areas like New York, Kansas City, and Rhode Island. Swift and Kelce's philanthropy mostly consists of food banks and education instituitions. Taylor Swift and Travis Kelce are marking one of the potentially biggest weeks of their relationship with a major philanthropic effort. Ahead of their expected wedding at New York's Madison Square Garden on Friday, the couple donated a combined $26 million to 20 local and national charities, according to Swift's publicist. The gifts were spread across organizations tied to places that have shaped their lives, from New York to Kansas City and Rhode Island. Nine of the nonprofits are based in New York, where the couple is expected to marry, including food relief organizations Food Bank For NYC and City Harvest, as well as Musical Mentors, which connects students with music teachers. Other recipients include the Rhode Island Community Food Bank, reflecting Swift's ties to Watch Hill, and Children's Mercy Hospital in Kansas City, where Kelce has spent his NFL career with the Kansas City Chiefs. National organizations such as Dolly Parton's Imagination Library, Feeding America, and the American Society for the Prevention of Cruelty to Animals also received donations. The amount awarded to each group was not disclosed. The donations continue a pattern of philanthropy from both stars. Swift has previously made multimillion-dollar gifts to food banks during her Eras Tour and has supported causes ranging from disaster relief to education. Kelce has long backed youth-focused initiatives through his foundation and has been recognized by the Chiefs for his charitable work. The pair, who began dating in 2023 and announced their engagement in 2025, have largely kept their wedding plans private. Charities react to the donations with gratitude For many nonprofits, the donations arrived as a surprise. City Harvest, New York City's largest food rescue organization, thanked Swift and Kelce on Instagram and wrote, "It's a love story that feeds good." The organization previously said that due to the Trump administration's cuts to the Supplemental Nutrition Assistance Program, it aims to deliver 100 million pounds of food to New Yorkers. In Kansas City, Harvesters — The Community Food Network said in an Instagram post that the couple donated $1 million. "The two of them have incredibly big hearts and have shown up to support not only us but many charities over the years," the organization wrote of Swift and Kelce. "We feel so lucky that they make the conscious effort to invest in hunger relief and the beautiful community we call Kansas City!" The Rhode Island Community Food Bank also posted on Instagram that Swift and Kelce surprised the organization with an "extraordinary" $1 million donation. "We were THRILLED to learn of this unexpected gift, which comes at a time when the need for food assistance in our state is at an all-time high," the Food Bank wrote. "Every dollar of this generous donation will be put to work to source food for distribution across our statewide network of 137 member agencies," the Food Bank added. Read the original article on Business Insider

Maggie Eastland / Bloomberg: Letter: chip sector group SEMI, including Micron and Samsung, warns Scott Bessent US policies affecting prices or production capacity would worsen the shortage  —  Government attempts to address the global memory chip shortage by influencing prices or production capacity would worsen …

Business Insider: Sources: Alexandr Wang said Meta's model currently in training, codenamed Watermelon, matches GPT-5.5 and uses an “order of magnitude more compute than Avocado”  —  Meta is making significant progress in the AI model race, its superintelligence chief Alexandr Wang told employees today.

The Information: Sources: Anthropic's bankers have hired UK law firm Freshfields to advise on its IPO; it also advised on Google's acquisition of Wiz and ServiceNow's Armis deal  —  A British law firm has scored a big win: a role in the Anthropic initial public offering, expected to raise tens of billions at a valuation of more than $1 trillion.

Discussion | Link

Nvidia CEO Jensen Huang has turned the black leather jacket into an instantly recognizable piece of tech-world iconography. I-Hwa Cheng/AFP; Sotheby's The leather jacket has become a staple of Nvidia CEO Jensen Huang's closet. Sotheby's is auctioning off one that was worn and signed by the Nvidia CEO himself. The proceeds will go toward grants, fellowships, and residencies for next-generation founders. Jensen Huang superfans: If you have at least $40,000 to spare, you have a shot at owning a valuable piece of hardware from the Nvidia CEO that isn't a chip. Starting July 7, Sotheby's will begin accepting bids on one of Huang's black leather jackets, a Tom Ford piece autographed by the CEO himself. The auction house estimates that the jacket will fetch $40,000 to $60,000. That's roughly the reported price of Nvidia's highly sought-after Blackwell AI chip. Jensen Huang was seen wearing this Tom Ford leather jacket at a tech conference in Taipei, according to an authenticator. Sotheby's A spokesperson for Sotheby's said the jacket will remain on display at Sotheby's New York until July 16. The sale will close on July 17. Like Steve Jobs with his signature black turtleneck, Huang turned his black leather jacket into a highly recognized piece of tech-world iconography as Nvidia has become a dominant force in the AI boom. The Nvidia CEO has been spotted in it for product launches and developer conferences, as well as on his trips abroad to countries like South Korea and Taiwan, where humidity makes a leather jacket a true commitment. Sotheby's said the jacket was photo-matched by Professional Sports Authenticator, a popular grading company, to the CEO's appearance at Hon Hai Tech Day in Taipei on October 18, 2023. Jensen Huang wore the Tom Ford jacket during his talk at Hon Hai Tech Day in Taipei. I-Hwa Cheng/AFP via Getty Images His signature was authenticated by James Spence Authentication, according to Sotheby's. The auction house said the sale was organized by Long Journey, an early-stage venture fund, for charity. The proceeds will go to a nonprofit called The Edge Institute and fund fellowships, grants, and residences for next-generation builders, the auction house said. Spokespeople for Tom Ford and Nvidia did not respond to a request for comment. Read the original article on Business Insider

Stephanie Stacey / Financial Times: Spotify removed 500K+ streams of Malcolm Todd's Earrings after a 70% surge in 24 hours sent it to #1 on Spotify USA and coincided with suspicious Kalshi wagers  —  Platform removes more than 500,000 streams of Malcolm Todd track ‘Earrings’ on concern traders propelled it to number 1