arXiv:2607.17106v1 Announce Type: new
Abstract: We propose a principled approach to reasoning about various graph algorithm implementations. We leverage the extended general Einsum notation (EDGE) which allows us to factor complexity along four axes: algebraic manipulation, mapping, format, and low-level implementations. Using breadth-first search (BFS) as a driving example and case study, we apply our methodology to explore over 90 variations across 26 categories of optimization choices for our GPU-based implementations. In addition to showing that our approach is general enough to represent previously discovered algorithmic techniques such as the pull variant of BFS, we discover novel variants that lead to geomean performance benefits ranging from 1.2x to 1.7x over the best Gunrock baseline variation for graphs with mid- to high- normalized degree variance.
Science Journals
arXiv:2607.16778v1 Announce Type: cross
Abstract: Treewidth is the standard measure for how ``tree-like'' a graph is. This paper studies how the treewidth of a product graph depends on the treewidth of its factors. Kozawa, Otachi, and Yamazaki [2014] and Hickingbotham and Wood [2025] independently showed that $\text{tw}(G\boxtimes H)\geq (\text{tw}(G)+1)\text{had}(H)-1$ for all graphs $G$ and $H$, where $\text{had}(H)$ is the Hadwiger number of $H$. We improve this bound to $\text{tw}(G\boxtimes H)\geq (\text{tw}(G)+1)(\text{tw}(H)+1)-1$, thereby solving an open problem of Hickingbotham and Wood. We also prove analogous product inequalities for pathwidth, Cartesian products, and strict bramble number, which is a parameter that is tied to treewidth. As an application of our results, we show that products of expanders have large subgraphs that are expanders.
arXiv:2607.16973v1 Announce Type: new
Abstract: Retrieval-Augmented Generation (RAG) systems increasingly power enterprise LLM applications, yet the vector retrieval layer introduces two underexplored challenges: (1) trained codebook quantizers may expose corpus statistics during index construction, creating a leakage channel in multi-tenant deployments, and (2) post-hoc filtering for tenant isolation degrades recall on selective queries. We study TurboVec, an open-source vector index built on TurboQuant - a codebook-oblivious scalar quantizer requiring no corpus-dependent training. On the DBpedia OpenAI embeddings benchmark (d=1536, 100K-999K vectors), TurboQuant 4-bit outperforms trained FAISS Product Quantization at the same memory budget by 8.5-8.9 percentage points in Recall@5 across all scales. Compared to HNSW (R@5=0.991) and IVF-PQ (R@5=0.840), TurboQuant occupies a distinct design point: higher recall than IVF-PQ without training, at 4-8x less memory than HNSW. Deployed on Snowpark Container Services, TurboVec achieves 11ms median query latency at 100K vectors versus 707ms for warehouse brute-force scan. Kernel-level allowlist filtering maintains 0.86-0.93 Recall@10 across 10-1000 tenant workloads versus 0.09-0.19 for post-filter baselines. Codebook-oblivious design reduces membership inference accuracy to near-random (50.0%) versus 57.3% for PQ codebooks. Limitations include single dataset evaluation, uncompressed HNSW comparison, and privacy evaluation on synthetic data only.
arXiv:2607.17538v1 Announce Type: new
Abstract: Retrieval-Augmented Generation (RAG) enhances the factual grounding of large language model (LLM) inference by retrieving relevant information from external knowledge bases. However, its dense vector retrieval introduces significant latency and energy overhead, becoming the primary performance bottleneck. Although recent in-storage accelerators aim to reduce data movement, they still rely on host or embedded processors outside the memory, where nearly 70% of the total retrieval time is spent. As a result, they cannot fully overcome the bandwidth limitations, leading to yet another memory bottleneck. To tackle these limitations, we present D-NOVA, a hardware-software co-designed in-storage retrieval accelerator. D-NOVA executes an inverted file (IVF)-based hierarchical retrieval pipeline by deeply embedding the search functionality directly into the NAND memory array. This is achieved by incorporating a new distance metric, Dual-Bound Tight Similarity Sensing (DTS), which is specifically tailored for searching within the NAND string. In addition, we introduce a lightweight contrastive adapter that maps embedding vectors into a DTS-friendly domain, recovering near-software recall while improving performance and energy efficiency. D-NOVA is up to 41.7x faster and 71x more energy-efficient than a CPU baseline, and achieves 12.13x higher throughput while being up to 1.26x more energy-efficient than state-of-the-art in-storage RAG accelerators, demonstrating the potential of fully in-storage vector search for scalable RAG acceleration.
arXiv:2607.16374v1 Announce Type: new
Abstract: Adversarial collaborations have become increasingly prominent in empirical science, where proponents of competing theories jointly design experiments capable of discriminating between them. Much less attention has been given to theoretical disputes that cannot readily be resolved by experiment. We distinguish theoretical adversarial collaboration from its empirical counterpart and propose a five-step template for conducting it. The procedure is designed to move disputants beyond familiar critique-reply-rejoinder exchanges by requiring sustained clarification, supervised steel-manning of the target theory, and the collaborative development and refinement of objections. Its condition for success is that the collaboration produce insights that substantially exceed the existing literature: either novel, non-obvious arguments show that the theory can meet the objections, or the theory's advocate concedes that new conceptual resources or revisions are required. The template was developed through reflection on, and implemented in, our own adversarial collaboration in philosophy and foundations of physics. We offer it here as a general framework for constructive and publishable progress in philosophy, foundational science, and other areas of inquiry in which disputes cannot readily be resolved by experiment.
arXiv:2607.16806v1 Announce Type: new
Abstract: Vision-Language Navigation in dynamic, human-centric environments exposes a fundamental tension: linguistic reasoning is slow and deliberative, whereas safe, socially compliant planning should be instant and reactive. The resulting observation staleness is safety-critical: a maneuver chosen during inference can already be unsafe by the time it executes. We observe that, long before a VLM finishes its inference, its intermediate hidden states already encode action-relevant intent. We propose SPARK-VLN, a dual-system framework for dynamic social VLN that streams the slow VLM reasoner's knowledge to a fast flow-matching expert planner throughout token generation, providing fresh and evolving guidance during inference. This design is realized by three modules: a Token-Wise Hidden Streamer that extracts intermediate hidden states along the token generation process, a Sequence-to-Slot Latent Bridge that projects them into fixed-size latent slots, and an Evolving Latent Conditioner that infuses them into the expert planner. We also introduce a human-centric benchmark suite for dynamic social vision-language navigation that keeps pedestrians and the robot active throughout inference and reports navigation success, social compliance, human collisions, and explicit staleness statistics. Across these settings, SPARK-VLN mproves navigation success and social compliance while sustaining inference efficiency. Webpage: https://hutslib.github.io/SPARK-VLN/.
arXiv:2607.17047v1 Announce Type: new
Abstract: LLM constraint reasoners are often evaluated near the random-SAT phase transition, confounding density and solver hardness. We test instance-level transfer while near-matching clause density. At aligned size bins, with near-matched density and matched maximum clause width, we compare proof-hard expander-Tseitin and proof-easy ladder-Tseitin formulas, pigeonhole anchors, and density-mismatched controls. Theory separates their resolution hardness; a solver-specific Glucose mean-conflict proxy differs by up to $51\times$, and five other solvers preserve the direction. Across three included models (243 instances each; a fourth is excluded for abstention), the near-matched-density accuracy gaps range from $-32$ to $+20$ points, with a pooled gap of $+1.7$ points ($p=0.74$) and a wrong-signed correctness-versus-conflict association ($r=+0.15$). A proof-preserving relabeling lowers accuracy in all five clusters for one model (mean $-93$ points) but not another, exposing model-surface sensitivity. In a preregistered extension, provider-reported completion-token spend does not consistently increase with the proxy after accounting for formula length and censoring. At 16k, the reasoning model spends more on proof-easy matched formulas and exhausts its budget on the solver-easiest UNSAT family; the 32k C1 gap is absent. These scoped dissociations concern verdict accuracy and observed token spend, not certificate solving, exact proof length, or allocation efficiency.
arXiv:2607.16580v1 Announce Type: cross
Abstract: Structural reliability analysis supports the design and safety assessment of buildings and civil infrastructure but requires specialized expertise throughout the workflow. This study presents a multi-agent large language model framework that automates component-level reliability analysis from a natural-language problem statement to interpreted estimates of the reliability index and failure probability. Specialized agents handle problem formulation, method planning, code generation, execution, and result interpretation, with human confirmation at key decision points. The Method Planner is fine-tuned using QLoRA for a priori reliability-method category selection. Analysis results are not generated directly by an LLM; instead, validated deterministic solvers compute the reliability estimates, improving reproducibility and reducing hallucination risk. The framework uses open-weight models and supports local execution without closed APIs. Results show that it lowers the expertise barrier to structural reliability assessment while preserving computational trustworthiness.
arXiv:2607.18051v1 Announce Type: new
Abstract: Detecting vessels engaging in illegal activities is of paramount importance for maritime security. One of the major goals is to detect dark vessels, ships that disable their transponders to evade surveillance. Deep Learning (DL) models can detect vessels in Synthetic Aperture Radar (SAR) images, enabling maritime traffic analysis regardless of weather or visibility conditions. However, to detect potential dark vessels, a DL model must select only those that are required to carry a transponder based on their Gross Tonnage (GT). Unfortunately, no public SAR dataset is available for training an end-to-end DL model for vessel detection and GT regression. In this work, we present a framework that leverages heterogeneous image and tabular datasets to solve this task. Our solution combines a multi-task DL framework for predicting the location, vessel type, and physical dimensions of ships, cascaded with a non-parametric model for predicting GT from vessel size and category. We perform GT regression by a KNN that measures sample similarity using a hybrid Euclidean and categorical distance. Experiments show that our solution can predict multiple outputs while remaining competitive with state-of-the-art models on individual subtasks, thus enabling the identification of dark vessels. We publish our code on GitHub https://github.com/PaltrinieriDavide/vesseldetection.
arXiv:2607.17052v1 Announce Type: new
Abstract: Vision-language models normally execute the same complete vision encoder for every question, even when OCR, counting, object, attribute, and spatial queries may not require identical computation. We study whether fixed-budget combinations of vision blocks can be skipped without fine-tuning. A shared K-block route skips one searched set of exactly K blocks for every question, while a capability-specific K-block policy selects one same-size route using a known capability label. We introduce a source-balanced evolutionary search and compare it with independent ranking, contiguous removal, and random routes at matched budgets. Experiments use Qwen2.5-VL-3B-Instruct, SmolVLM2-2.2B-Instruct, and an 876-example image-disjoint selection split. Search transfers across architectures: on SmolVLM2, the searched shared four-block route beats independent construction by 4.91 percentage points. Capability specialization is less stable. On Qwen, the six-block capability policy beats the shared route by 2.17 points, driven by a 7.10-point OCR gain. On sealed IIIT5K, however, the SmolVLM2 OCR-specific route trails its shared route by 13.6 points. Combinatorial search reliably improves route construction, but capability labels do not define universally transferable vision pathways.
arXiv:2607.17111v1 Announce Type: new
Abstract: We present the Bangladesh Legal Acts Dataset (BLAD), a curated collection of 1{,}484 legislative acts enacted between 1799 and 2025. Each act is represented with its full text, structured sections and footnotes, repeal status, and metadata linking it to the governing regime, head of state, and prevailing legal framework at the time of enactment. The corpus spans English, Bengali, and mixed-language documents, supporting temporal and multilingual analysis of statutory law. BLAD addresses a persistent gap in legal natural language processing (NLP) resources for low-resource, civil-law jurisdictions in South Asia. We describe the acquisition and enrichment pipeline, report descriptive statistics over more than two centuries of legislation, and outline the research directions the corpus enables. The dataset is publicly available under the CC~BY-SA~4.0 license at https://www.kaggle.com/datasets/sakhadib/bangladesh-legal-acts-dataset.
arXiv:1905.12913v2 Announce Type: replace
Abstract: This paper investigates the problem of utilizing network topology and partial timestamps to detect the information source in a network. The problem incurs prohibitive cost under canonical maximum likelihood estimation (MLE) of the source due to the exponential number of possible infection paths. Our main idea of source detection, however, is to approximate the MLE by an alternative infection path based estimator, the essence of which is to identify the most likely infection path that is consistent with observed timestamps. The source node associated with that infection path is viewed as the estimated source $\hat{v}$. We first study the case of tree topology, where by transforming the infection path based estimator into a linear integer programming, we find a reduced search region that remarkably improves the time efficiency. Within this reduced search region, the estimator $\hat{v}$ is provably always on a path which we term as \emph{candidate path}. This notion enables us to analyze the distribution of $d(v^{\ast},\hat{v})$, the error distance between $\hat{v}$ and the true source $v^{\ast}$, on arbitrary tree, which allows us to obtain for the first time, in the literature provable performance guarantee of the estimator under limited timestamps. Specifically, on the infinite $g$-regular tree with uniform sampled timestamps, we get a refined performance guarantee in the sense of a constant bounded $d(v^{\ast},\hat{v})$. By virtue of time labeled BFS tree, the estimator still performs fairly well when extended to more general graphs. Experiments on both synthetic and real datasets further demonstrate the superior performance of our proposed algorithms.
arXiv:2607.17328v1 Announce Type: new
Abstract: Traditional time-lock puzzles enforce delayed access to encrypted secrets by relying on inherently sequential computational steps. However, since software-based constructions impose no fundamental bound on the physical execution speed of individual steps, they remain vulnerable to hardware acceleration and improved implementations. In practice, this renders existing schemes ``step-lock'' rather than true ``time-lock'' puzzles, making long-term delay guarantees highly speculative against decades of unpredictable hardware advancement. To address this fundamental limitation, we introduce the Physical Time-Lock Puzzle (P-TLP), a new paradigm that anchors solving delay directly to the intrinsic, hardware-bounded evaluation latency of silicon hardware. By leveraging noisy Physical Unclonable Functions (PUFs) as non-parallelizable delay oracles, P-TLPs establish resistance against both parallel computation and algorithmic acceleration under the random oracle model. To tolerate PUF evaluation noise while maintaining a tightly concentrated and predictable solving-delay window, we propose a composite puzzle architecture that combines multiple independent basic puzzles. We formally prove the optimality of a sequential greedy solving strategy and derive tightly concentrated solving-delay windows using Hoeffding and Berry-Esseen bounds. We additionally prove early-solve hardness, establishing that adversaries with substantial classical computation budgets cannot bypass the physically-enforced PUF evaluation bottleneck. We validate our theoretical analysis with an FPGA prototype built around a configurable Ring Oscillator PUF, confirming that our analytical delay predictions tightly match empirical measurements. Our results demonstrate that P-TLPs can deliver highly predictable, long-term delay guarantees, enabling practical deployment in high-value applications such as digital legacy management.
arXiv:2607.17090v1 Announce Type: cross
Abstract: We study the expressivity of shallow polynomial neural networks (PNNs) with monomial activation functions over finite fields. For a given architecture, we define a neuromanifold as the image of the map from all possible network weights into the product of polynomial rings. We quantify the expressivity by the cardinality of the neuromanifold, and derive a natural lower and upper bound. This leads to counting rational points over finite fields, a problem closely linked to the Weil conjectures. Finally, we present an architecture that exhibits a striking difference in the neuromanifolds when considered over a characteristic zero versus a finite-characteristic field, illustrating the critical role of field characteristic in the notion of expressivity.
arXiv:2607.16378v1 Announce Type: new
Abstract: Estimating the apparent age of individuals from facial images is challenging due to the subjective nature of perception and the inherent variability of the data. We investigate the role of uncertainty estimation, attributing uncertainty jointly to a lack of knowledge (epistemic) or inherent noise/chance (aleatoric). Leveraging the APPA-REAL dataset, we train Bayesian Neural Networks on datasets of varying sizes using three BNN approximations: MC-DropConnect, Flipout, and Deep Ensembles using supervision on human aleatoric uncertainty available in the APPA-REAL dataset. Each model outputs both the predicted apparent age and the amount of aleatoric and epistemic uncertainty. Our results confirm the hypothesis that the inherent aleatoric uncertainty remains stable across dataset sizes, while epistemic uncertainty increases as training data decreases. These findings demonstrate that different sources of uncertainty can be quantified in face age estimation.
arXiv:2607.16379v1 Announce Type: new
Abstract: One of us is a realist and believes that reality fundamentally consists in an external physical world, governed by laws of physics. Observers are either emergent (physicalism) or exist in addition to physical reality. The other defends a version of idealism and believes that reality fundamentally consists in first-person states, unembedded into worlds, on which laws of nature act directly. Shared "physical worlds" are merely emergent. The disagreement ultimately centers on the overall coherence of a formal model, known as Algorithmic Idealism, and its ability to resolve observer paradoxes, such as the Boltzmann brain paradox. Our aim is to confront this issue in the form of an adversarial collaboration, where we avoid misunderstanding each other, so that readers can see the true source of the disagreement, and decide for themselves. Our work represents an attempt to establish a form of exchange that may help overcome the fragmentation of scholarly communities in philosophy, physics, and elsewhere. The collaboration was conducted using a discipline-neutral template for theoretical adversarial collaboration developed in a companion paper [1], which we hope will be useful in other theoretical disciplines.
arXiv:2607.18052v1 Announce Type: new
Abstract: Over-the-air computation (OAC) enables efficient function aggregation in wireless networks by exploiting the superposition property of the multiple-access channel. However, practical deployment of OAC is severely challenged by the reliance on accurate carrier synchronization and coherent reception, which are costly and fragile, especially in short-range and low-complexity systems. In this work, we propose a \emph{self-coherent, synthesizer-free over-the-air computation framework} based on \emph{Kramers--Kronig (KK) reception}. By transmitting a biased aggregate waveform and employing direct detection followed by KK phase reconstruction at the receiver, the proposed scheme eliminates the need for explicit carrier recovery while preserving coherent-like signal aggregation.
We develop a signal-domain system model for multi-user OAC under KK reception and provide a synchronization-relaxation analysis demonstrating that the proposed architecture fundamentally removes carrier-frequency offset (CFO) sensitivity between transmitters and receiver. By shifting synchronization complexity away from strict carrier-phase tracking and eliminating distributed phase alignment requirements, the framework reduces control overhead and improves scalability in multi-user aggregation. A detailed per-symbol mean-squared error (MSE) characterization isolates the impact of channel mismatch and KK reconstruction noise, showing that the proposed self-coherent architecture approaches the theoretical performance limits of baseband OAC under practical operating conditions. Finally, we demonstrate that the approach is particularly well suited for mmWave and sub-THz systems, where oscillator phase instability otherwise represents a fundamental bottleneck to scalable coherent OAC.
arXiv:2607.18056v1 Announce Type: new
Abstract: Frontier large language models (LLMs) are increasingly integrated into scientific workflows, yet their growing biological capabilities may outpace current safeguards. To assess the biological risks of frontier models, we develop Intern-BioBreaker, a specialized bio-red-teaming model, together with an integrated computational-to-physical framework that couples model-level stress testing with wet-lab validation. Within this framework, Intern-BioBreaker generates targeted jailbreak prompts to test whether aligned models can be induced to provide operational guidance for safety-sensitive biological tasks or produce sequence-level outputs with potentially harmful properties. Selected sequence outputs are then carried forward for DNA synthesis, host expression, and orthogonal protein verification to assess whether model-generated designs can yield the intended biological products. Our evaluation reveals a concerning gap between text-level safeguards and the risks posed by capable scientific models: (i) Intern-BioBreaker outperforms baseline attack models and reveals widespread bio-risk jailbreak vulnerabilities across both open-weight and proprietary frontier LLMs, with several targets reaching near-saturated or 100% task-level attack success rate (ASR); (ii) in sequence-level case studies, GPT-5.5 can be induced to generate modified viral candidate sequences with pathogenic potential; the corresponding translated proteins may exhibit even stronger receptor-binding affinity and thus enhanced infection potential; and (iii) end-to-end verification shows that selected model-generated biological designs are not merely textual artifacts, but can be physically realized under controlled experimental settings. These findings underscore the need for stronger biological red-teaming, nucleic acid synthesis screening, and safety mechanisms that keep pace with model capabilities.
arXiv:2607.16383v1 Announce Type: new
Abstract: Verifying academic credentials remains difficult: records are held by individual institutions in proprietary systems, verification is slow and manual, and counterfeit qualifications are widespread. Blockchain-based registries have been proposed as a remedy, but existing systems tend to anchor certificate hashes without binding them to a verifiable identity, without an explicit mechanism to accredit issuing institutions, and without support for correcting or revoking credentials once issued. This paper investigates whether an infrastructure designed for regulated financial instruments can be repurposed to close these gaps. We present the design of a registry for identity-bound academic credentials that composes OnchainID self-sovereign identities (ERC-734/ERC-735) with the T-REX suite (ERC-3643): its trusted-issuer registry becomes an on-chain issuer-accreditation whitelist, and each certificate is represented as a signed, updatable claim bound to a student's identity and verifiable by any third party without a wallet, while sensitive fields are kept off-chain. We make explicit the tension between a transferable-security-token standard and non-transferable credentials, clarifying which of its guarantees carry over. We validate the design with a reference implementation covering the full certificate life cycle and evaluate it in terms of gas cost, scalability, latency, and security, quantifying the overhead relative to a hash-anchoring baseline.
arXiv:2607.16540v1 Announce Type: new
Abstract: Understanding and characterizing human color perception is a longstanding research goal. One of the most traditional approaches is looking for the human color discrimination thresholds, the minimum chromatic differences perceptible to human observers. In recent years, deep neural networks have become the standard networks for computer vision tasks. In particular, deep vision encoders, foundation models trained on large-scale visual data, map images into latent feature representations. Despite the widespread use of deep vision encoders, few studies have investigated whether their internal representations exhibit human-like discrimination thresholds. In this work, we present a large-scale exploratory study probing the chromatic sensitivity of more than 50 pretrained vision encoders, including convolutional networks and vision transformers, against human discrimination thresholds. Using controlled chromatic stimuli at multiple chroma levels, we compare model-derived chromatic discrimination thresholds with human discrimination ellipses through a region-overlap metric (mIoU). Our analysis reveals generally weak alignment between model representations and human perceptual thresholds across all model families, with the best mIoU < 0.25. Moreover, we find that self-supervised encoders consistently outperform supervised ones, while language-supervised models show the most polarized behavior, occupying both the top and bottom of the ranking. These findings suggest that human-like chromatic sensitivity does not emerge naturally from current large-scale visual training objectives for any of the analyzed architectures.
arXiv:2607.18058v1 Announce Type: new
Abstract: Although a complete characterisation of the probability distribution in the phase space of turbulent flows remains elusive, accurately sampling this distribution is essential for both synthetic turbulence generation and turbulent flow reconstruction. Motivated by these applications, we examine to what extent a machine-learned distribution can approximate the physical invariant distribution of turbulent channel flow at $\mathrm{Re}_\tau=180$. We assess three important properties of the approximation: physical ensemble statistics, consistent conditional sampling, and dynamical invariance. To this end, a flow-based generative model is trained on a minimal conditional flow unit, which we define as the smallest domain outside which conditional fields, given a single observation at the domain centre, are indistinguishable from unconditional fields in terms of mean-square discrepancy to other conditional fields. We also introduce a consistent procedure for sampling from the conditional learned distribution. Comparisons with direct numerical simulation show that synthetic turbulent fields reproduce key statistical and dynamical features of turbulence, including intermittency and nonlinear energy transfer. The consistency of conditional sampling is demonstrated in a flow reconstruction problem, and subsequently used to generate synthetic turbulent velocity fields on a large domain. When adopted as initial conditions in direct numerical simulations, these fields yield physical and statistically stationary ensemble statistics, indicating that the learned distribution provides a good approximation to the natural distribution of the turbulent dynamical system.
arXiv:2607.18060v1 Announce Type: new
Abstract: Long-horizon robotic tasks require diverse capabilities that no single policy can reliably provide. Heterogeneous policies offer complementary strengths, but orchestrating them requires reasoning over uncertain capability boundaries and cross-policy distribution mismatch, which are largely overlooked by existing planning methods built on homogeneous, predefined skills with fixed applicability. We propose RoboHarness, a unified framework that encapsulates independently developed robot control systems as reusable agentic skills. Although instantiated in this work with VLAs, RL policies, and task-and-motion planning (TAMP) systems, RoboHarness is designed as a general framework compatible with a broader range of robot policies, such as navigation policies, model predictive controllers, and world-action models. RoboHarness uses multi-modal execution memory and online evidence to characterize policy capability boundaries for capability-aware decomposition and routing. To stabilize policy handoffs, its Memory Bridge retrieves execution trajectories associated with the next policy, estimates its in-distribution state region, and guides the robot toward that region without joint policy retraining. Extensive experiments on three public benchmarks, 500 customized tasks, and 135 real-robot experiments demonstrate effective capability-aware routing and stable policy orchestration, yielding substantial improvements in zero-shot long-horizon planning and out-of-distribution robustness.
arXiv:2607.16979v1 Announce Type: new
Abstract: Repository mining studies increasingly analyze AI-evidence projects, yet it remains unclear how to measure whether architectural changes create deferred stabilization obligations. A natural metric, post-anchor stabilization density, counts tests, CI gates, documentation, and fixes appearing after a durable boundary is introduced. We show that this metric fails. In a diff-level study of 338 high-visibility 2026 GitHub repositories, anchors are common (321 of 338 contain real changed-file anchor evidence), but a controlled 308-event contiguous-window experiment finds no post-anchor uplift: broad and strict stabilization signals both yield median post/pre density ratios near 1.0, and non-anchor controls are equally dense. We introduce stabilization regimes, six recurring patterns that explain why the density metric fails, and use human validation to calibrate them. Two independent coders label 100 stratified candidate-anchor events from blinded packets (kappa = 0.50 on debt attribution). The validation exposes a two-layer trap: many candidate anchors are not durable boundaries (45 of 100), and even among valid anchors in this calibration sample the no-uplift result holds: only 3 of 100 events survive as attributable delayed obligations; the remaining 97 are explained by classifier error, anchor-local hardening, background maintenance, or pre-anchor hardening. Post-anchor density conflates pervasive maintenance with genuine debt; controlled designs with regime-aware attribution are necessary before repository mining can reliably identify stabilization obligations.
arXiv:2607.17593v1 Announce Type: new
Abstract: Class Incremental Learning (CIL) aims to learn new concepts consistently from a data stream without forgetting. Unlike typical CIL methods which need to learn a model from scratch, pre-trained model (PTM) can easily adapt to a new task with fine-tuning. However, existing PTM-based CIL methods fail to achieve a trade-off between performance and computational expenditure, i.e., they either adopt the same parameter space so that leading catastrophic forgetting, or expand a new branch for each task but adding more computational cost. To this end, we propose MetrIc Learning with Expandable Subspace (Miles) to harness the prior information within pre-trained knowledge, thereby orchestrating an efficient expansion of the parameter space through guided optimization. Specifically, it decouples the learnable modules with the pre-trained model, exploiting prior information from intermediate features of the backbone network to enable more flexible parameter expansion. Then, a central loss is adopted to guide the new category to cluster towards the corresponding prototype in the new task subspace while incorporating an auxiliary distance regularization term to maintain metric equilibrium across tasks. Extensive experiments on six benchmark datasets demonstrate that Miles achieves state-of-the-art performance in various CIL settings.
arXiv:2607.17329v1 Announce Type: new
Abstract: Medical image segmentation models require both high accuracy and lightweight design to accommodate real-world medical applications. The deployment of these models on resource-limited medical platforms remains a significant challenge due to their high computational and parameter requirements. Existing pruning methods for model compression mostly overlook the intrinsic connections and similarity between the internal structures of complex deep neural networks. As a result, compressed models may not effectively retain the basic features of the pretrained network. To solve this problem, we propose a hierarchical clustering compression method for medical image segmentation models (MIS-HCC). This approach employs hierarchical clustering to partition channels and fuse their parameters efficiently. Specifically, it leverages the Wasserstein distance to represent similarity of channels within layers of pre-trained network, forming a similarity matrix that guides the clustering process. Channels within each cluster are then fused to produce a compressed network. Experimental results on three medical image datasets application demonstrate that MIS-HCC outperforms the state-of-the-art methods in both accuracy and compression efficiency, offering an effective solution for deploying medical image segmentation models on resource-limited medical platforms.