arXiv:2607.18013v1 Announce Type: new
Abstract: This paper introduces a method for real-time processing and transmission of autonomous underwater vehicle (AUV) imagery over low-bandwidth communication links. It leverages artificial intelligence (AI) techniques to identify a set of images that best represent an entire dataset, or automatically finds the most similar images to a given query image for transmission to operators. Combined with metadata of a larger set of images, compressed versions of the selected images can be transmitted over satellite communication links or underwater modems, and provide operators on shore with information about the type of imagery the AUV is collecting while it is still deployed. Data from three deployments off the coast of the UK and in Gran Canaria using different AUVs and imaging systems demonstrate the method in the field. It achieved an almost 400,000-fold reduction in data volume compared to the raw data size, enabling transmission of data summaries of a 2-hour 47-minute-long mapping mission in just over 34 minutes over low-bandwidth satellite communication.
Science Journals
arXiv:2607.16845v1 Announce Type: new
Abstract: Scientists at European XFEL conduct experiments that generate very large and complex datasets. The subsequent data analysis is challenging as scientists must combine their domain expertise with facility- and software-specific knowledge scattered across documentation, tools, and support channels. To address this problem, we designed and evaluated an agentic AI system tailored to the scientists' needs and integrated with the high-performance computing environment of European XFEL. Using a design science research approach, we conducted a rapid literature review, a systematic evaluation of 16 AI tools, multiple interviews, a focus group, and a user study with experts at European XFEL to develop and evaluate two prototypes. Our study identifies key knowledge challenges in scientific data analysis, derives requirements for an AI agent that supports knowledge retrieval and source code generation, and proposes design recommendations for a specialized system adaptable to the evolving AI tool landscape. These findings provide guidance for developing maintainable AI support in highly specialized scientific environments.
arXiv:2607.17445v1 Announce Type: new
Abstract: Non-interference properties, spanning confidentiality and integrity, have long enjoyed a position as the high water mark of program security guarantees. Information flow type systems comprise the primary means for obtaining non-interference properties of programs, but their potential as a holy grail for secure programming has remained latent. Prior work bifurcates the type system along confidentiality and integrity, resulting in duplicate reasoning machinery and complex specifications. Furthermore, long-held wisdom dictates that non-interference must be weakened with downgrading mechanisms to accommodate the needs of practical programs, nearly all of which violate confidentiality and integrity in the course of fulfilling their purpose. This often pierces abstraction barriers and compromises modular reasoning.
We introduce parametric information flow, which uses recent insights from modal type theory to shed light on these issues. In particular, we draw inspiration from work on the open and closed modalities, highlighting their rich interplay. Though each individual modality finds uses throughout the literature, our key insight is that their joint interaction suffices to reconstruct full-spectrum information flow reasoning, producing a single framework accounting for both confidentiality and integrity. Downgrading and analogues of advanced reasoning tools in the lineage of robust declassification are recovered without extensions to our theory, strengthening prior results. We show non-interference via a binary logical relations argument, realizing robustness as an ordinary 2-hyperproperty mediated by our modalities. Our work reveals state-of-the-art downgrading mechanisms to be wholly compatible with those for abstraction and modularity, arising precisely from the semantics of the latter under full-strength non-interference.
arXiv:2607.17523v1 Announce Type: new
Abstract: Recent advances in world models and video generation have given rise to an emerging reasoning paradigm that leverages video generative models to simulate, predict, and reason about real-world dynamics. We redefine this paradigm as Thinking in Video, where video is not merely an output artifact but a medium for constructing, extending, and verifying causal thought. However, this promise remains unverified: convincing rollouts may reflect memorized appearances rather than causal understanding, while existing metrics separate perceptual fidelity from semantic logic. To evaluate whether video generators support such reasoning, we introduce the Causal-Generative Dual-Judge (CGDJ), auditing World Model Consistency from two perspectives. Explicit Causal Perception tests whether a generator reads a video scenario as a reasoning problem through spatio-temporal flattened visual question answering, while Implicit Generative Perception-Prediction Gap evaluates whether it renders the causal consequence as a consistent future video. Applying CGDJ to representative open- and closed-source generators reveals a clear Perception-Prediction Gap: open-source models produce plausible dynamics despite near-zero explicit causal perception, whereas advanced closed-source systems show stronger but still limited alignment between reasoning and generation. Further analysis exposes audio-visual misalignment, where models verbalize correct causal logic more reliably than they render it, challenging the "world simulator" narrative.
arXiv:2607.17864v1 Announce Type: new
Abstract: Global linear stability analysis of bluff body wake flows is performed using the matrix-forming method based on finite-difference discretization. Particular emphasis is placed on the influence of outflow boundary conditions, with the aim of minimizing the required computational domain size without degrading accuracy or inducing spurious oscillations near the outlet. This study focuses on incompressible wakes behind bluff bodies such as cylinders and airfoils at high angle of attack, especially in regimes where global modes exhibit downstream spatial amplification. It is shown that below the critical Reynolds number -- where the global mode remains linearly stable -- significant spatial growth can persist far downstream, even when the wake is nearly absent. This behavior underscores the importance of imposing a physical boundary condition at the outlet. Several commonly used outflow boundary conditions are evaluated, including Dirichlet, Neumann, extrapolation, stress-free, sponge layer, and the Robin condition that incorporates predictions from local linear stability analysis at the outlet. The results demonstrate that, for different $Re$ cases, the Robin condition enables robust convergence of global modes within substantially truncated domains, thereby improving the efficiency of global stability analysis. These findings highlight the broader applicability of the matrix-forming approach for complex stability analyses, including Floquet analysis of time-periodic flows and extensions to compressible configurations.
arXiv:2506.04122v3 Announce Type: replace
Abstract: Open-loop performance evaluation of 3D multi-object tracking in autonomous driving requires matching criteria that effectively penalize translational, shape, and orientation errors from the ego vehicle perspective. The prevailing criteria for determining true positives are Intersection over Union (IoU) and Centre-Point Distances (CPD). When IoU is extended from the 2D image plane to 3D volumetric overlap, it often falls below its acceptance threshold even with minor yaw misalignments, whereas CPD disregards orientation entirely. To address this limitation, we propose Contour Errors (CE) as an ego-centric criterion that employs Hausdorff-type reasoning to sparse bounding-box corner geometry by selecting the k-nearest ego-centric corners. This method provides a graded orientation sensitivity between the extremes of IoU, which overpenalizes, and CPD, which is orientation-blind. We evaluate Contour Errors against six baselines using the HOTA evaluation protocol on the nuScenes dataset, conditioned on proximity, yaw error, and a confidence threshold. At the standard IoU vehicle threshold, 47% of car and 75% of pedestrian CE-valid matches are rejected by IoU despite close contour proximity, while fewer than 0.1% of IoU-valid matches fail CE, resulting in a 4.2% higher match rate relative to total ground truth annotations. These results establish the ego-centric matching criterion as a primary factor for improving open-loop perception evaluation in safety-critical autonomous driving.
arXiv:2607.18228v1 Announce Type: new
Abstract: To test how correct logical judgments respond to learned context, we prepend a soft prefix to an exactly labeled syllogistic reasoning benchmark while keeping the model fixed. Soft prefixes are opaque continuous vectors, so we characterize them through the behavior they induce across controlled variations in logical form and interface. By studying which prefixes succeed and how their effects generalize, we characterize how learned contextual pressure can override correct judgments and expose limits in a model's logical stability. Across Qwen3.6-35B-A3B MoE, Qwen3-8B, and Gemma 4 31B, learned prefixes redirect many correct answers and remain effective across unseen forms and interface changes. In repeated tests with Qwen3.6 MoE and Gemma, they outperform paired random controls in all 16 model--direction--split comparisons by 37 to 99 percentage points. Qwen3.6 MoE flip rates remain between 72% and 90% across wording and prompt changes, while Gemma validity prefixes retain 54% to 56% flip compared with less than 1% for matched random prefixes. Diagnostic tests show that the dominant effect is a broad preference for one answer meaning rather than fixed-symbol forcing or a logical operation that transfers reliably between tasks. The form of this bias differs across models. In both Qwen models, simple score models often predict which judgments will flip but not how far their margins will move, whereas Gemma's overall response is more closely approximated by the same models. These results show that the dominant behavioral effect of successful soft prefixes is a broad answer preference, while the remaining response reveals substantial model-specific differences in logical stability.
arXiv:2607.18230v1 Announce Type: new
Abstract: Modern vision-language models (VLMs) have significantly improved image generation and editing capabilities, making pixel-level image tampering detection increasingly important yet challenging under cross-model and out-of-distribution shifts. This work studies domain generalization for pixel-level image tampering detection in modern VLMs like ChatGPT, Gemini, Qwen-Image, etc., aiming to learn tampering localization models that remain robust across diverse VLM-generated manipulation distributions. We propose a simple yet effective domain-generalized training framework built on two practical strategies. First, we introduce a balanced minibatch sampling scheme that strategically samples tampered and real images in each minibatch, preventing biased optimization toward either manipulated artifacts or clean-image priors and avoiding training collapse, ensuring that each optimization step receives proper sampled gradient signals. Second, we adopt a simple late-injection strategy, where the detector is first trained on large-scale base data until stable convergence, and then exposed to a small amount of newly selected supporting data from emerging VLM distributions, improving adaptability without overfitting to limited new domains. Together, these components provide a simple yet strong recipe for improving pixel-level tampering localization and OOD robustness across modern VLMs. Despite the conceptual simplicity, our framework outperforms the prior state-of-the-art PIXAR by a large margin of 26.1% and 26.8% relative improvement in average gIoU and cIoU, respectively, across OOD VLMs of GPT-Images-2.0, Gemini-3.1, FLUX.2, and Seedream 4.5. Our code is available at https://github.com/VILA-Lab/PIXAR-DG
arXiv:2607.17132v1 Announce Type: new
Abstract: Digital twins enable robots to anticipate and adapt to physical interactions, but existing models struggle with elastoplastic articulated objects (EAOs) that exhibit nonlinear elasticity, plastic yielding, and damage accumulation. We present BoxTwin, an interactive digital twin framework that learns the full dynamics of EAOs from videos. Our pipeline reconstructs the scene, identifies a physics aware constitutive model for each EAO. Experiments on manual folding and dual arm manipulation of EAOs show that BoxTwin accurately tracks joint trajectories and reproduces post contact plastic behavior over long horizons. By integrating video driven reconstruction with elastoplastic damage modeling, BoxTwin advances digital twins toward predictive, adaptive control of deformable articulated objects in unstructured environments.
arXiv:2607.17473v1 Announce Type: new
Abstract: Click-through rate (CTR) and conversion rate (CVR) prediction are fundamental tasks in online advertising, aiming to estimate the likelihood of user interactions based on various features. While personalized attributes such as age and gender can significantly enhance predictive accuracy, their use is increasingly restricted by privacy regulations, thereby limiting available data for both training and inference. To address this challenge, we propose RAMP (Robust Ad Recommendation Under Limited Personalized-Feature Availability via Masking and Alignment Pathways), which is designed to improve CTR/CVR prediction accuracy when personalized features are not accessible, thus supporting deployment in privacy-constrained settings.RAMP consists of (i) a personalized pathway built upon a dual-tower component with identical inputs but independent parameters, where output masking separates predictions for personalized and non-personalized signals, (ii) a separate non-personalized pathway trained with non-personalized features only, and (iii) a distillation-inspired prediction-alignment architecture between (i) and (ii) that improves prediction when personalized features are unavailable. We conduct comprehensive experiments using both public benchmarks and industrial datasets to evaluate the performance of RAMP. Our evaluation spans multiple backbone models and different settings: with and without access to personalized features. The results show that RAMP consistently outperforms state-of-the-art methods when personalized features are missing, while maintaining competitive performance when all features are available. %demonstrating its effectiveness and practicality for real-world advertising systems. Our code is publicly available at https://github.com/Ruixinhua/RAMP.
arXiv:2607.17401v1 Announce Type: new
Abstract: Super-resolution can make inspection images appear sharper without preserving the evidence needed to detect a defect. We study this failure mode with a benchmark that separates reconstruction from detection and evaluates both at a predeclared low false-positive rate. Ten end-to-end repetitions combine independently generated line/space and contact-hole images with model training, calibration, clean controls, weak defects, and a held-out defect morphology. Every reconstruction is scored by the same local residual detector, while direct and jointly trained detectors form a separate comparison track. Reconstruction fidelity and inspection utility diverge: the two learned reconstruction models attain the highest structural similarity yet detect fewer defect pixels than bicubic interpolation in every paired repetition. A direct DeepLabV3 detector reaches $0.1984\pm0.0385$ recall at $0.000174\pm0.000084$ false-positive rate and satisfies the held-out feasibility criterion in all ten repetitions. An illustrative joint model, DPU-WaferSR, passes independent clean calibration but exceeds the held-out limit in all ten repetitions, demonstrating that calibration success does not guarantee transfer. Weak-defect recall remains near zero for every feasible method. Applying the unchanged policies to 4,591 public Carinthia-S masks further reveals large method-dependent shifts on real SEM texture. These results support a simple conclusion: super-resolution for inspection should be judged by preserved task evidence and operating-point transfer, not reconstruction quality alone.
arXiv:2607.17268v1 Announce Type: new
Abstract: PairUAV relative localization maps two UAV images to a polar navigation command. Although heading and range share the same pairwise pose context, treating them as homogeneous coordinates forces both outputs to use the same decoder evidence and optimization state. Controlled readout probes reveal a different structure: the two axes favor different decoder-depth combinations, their best checkpoints disagree on 80.8% of a validation trajectory, and range errors exhibit a distinct high-error tail. We introduce method, Polar Axis-Conditioned Estimation, which retains a shared Reloc3r-style pair representation while assigning axis-specific readout interfaces. Heading uses mid/late relational evidence, whereas range remains attached to a direct late metric path. On the official hidden test, the strongest released raw predictor scores 0.002460; the complementary PAAER predictor scores 0.002514 with a slightly lower angle error. Deterministic challenge packaging, reported separately from learned estimation, yields the final score of 0.001874. Code, checkpoints, predictions, and reconstruction tools are available at https://github.com/zerong7777-boop/PairUAV-PACE.
arXiv:2607.16864v1 Announce Type: new
Abstract: Supercharging of lithium-ion batteries (LiBs) requires robust health monitoring to ensure durability, safety, and user confidence, particularly for emerging vehicle-to-grid applications with bidirectional energy flows. Yet battery management remains largely disconnected from the material and structural origins of aging, limiting both interpretable health assessment and informed battery design. Here we propose a physics-informed learning framework with virtual sensing that infers hard-to-measure design parameters, including solid-state diffusion coefficient, electrode thickness, ion concentration, and particle size, directly from standard battery management system (BMS) measurements. Across diverse fast-charging strategies and driving profiles, embedding a digital-twin-derived particle-cracking mechanism as a soft constraint reduces trajectory and lifetime prediction errors by 6-8 times relative to state-of-the-art machine learning baselines using only 2% early-life observations. We further show that accurate degradation extrapolation does not require fully resolved governing equations; validated partial mechanisms, jointly refined with limited data, provide sufficient guidance. Virtual sensing transforms standard charging signals into latent design variables without additional sensors, bridging observable battery behavior and underlying aging processes while reducing capacity loss error by up to 39%, end-of-life (EOL) error by 17%, and prediction variability by up to 54%, enabling real-time exploration of new battery configurations. More broadly, the proposed framework establishes a practical feedback loop between deployment and development, demonstrating how real-world operation can continuously inform upstream design decisions across complex multiphysics systems.
arXiv:2607.17537v1 Announce Type: new
Abstract: Spin-phonon coupling governs magnetic relaxation in numerous systems including single-molecule magnets and molecular spin qubits. In most cases, the accurate prediction of spin relaxation rates requires multireference electronic structure methods, but their computational cost has largely restricted such calculations to isolated molecules. Here we show that spin-phonon relaxation rates can be computed within a multireference density matrix embedding framework. We apply the approach to three cobalt- and two dysprosium-based single-molecule magnets and to a cobalt-based molecular crystal. Across all systems, treating only the first coordination sphere of the magnetic center at the multireference level reproduces spin relaxation rates in good agreement with non-embedded CASSCF calculations while reducing the correlated problem to 10-68% of the total basis functions. Periodic calculations further demonstrate that spin relaxation rates can be computed for a molecular crystal using an embedded active space of only 259 basis functions out of a total of 2569. These results show that multireference density matrix embedding extends quantitative spin-phonon relaxation calculations from isolated molecules to molecular crystals.
arXiv:2607.16799v1 Announce Type: new
Abstract: Hidden-state readouts can predict whether language-model outputs are correct, but objective correctness (OC) usually agrees with the model's own self-judgement (SJ), leaving the decoded signal semantically ambiguous. We construct conflict cases in which OC and SJ predict opposite readout orderings. On high-confidence disagreements, conventional correctness-labelled contrasts often rank incorrect/self-endorsed responses above correct/self-rejected responses, following SJ rather than OC. We estimate factorial SJ- and OC-associated directions and evaluate their polarity across mathematical reasoning and factual recall. Across four instruction-tuned models up to 14B parameters, the SJ-associated direction transfers above chance in both cross-domain directions for every model, whereas the OC-associated direction has a below-chance point estimate for the expected OC ordering in every corresponding condition. This transfer asymmetry develops across middle-to-late layers, persists under answer-likelihood, sequence-length, and null-direction controls, and extends to MMLU and binary TruthfulQA without target-domain direction fitting. Across the studied models and diagnostic subsets, the most reliably transferable component preserves SJ-associated polarity. Transferability alone therefore does not establish objective-correctness semantics.
arXiv:2607.16801v1 Announce Type: new
Abstract: Simulating dependent Bernoulli outcomes with prescribed means and pairwise Pearson correlations is a common task in risk modeling. A familiar approach is the Gaussian-threshold workflow for binary outcomes, often viewed as a Bernoulli analogue of the Gaussian copula construction. We show that setting latent Gaussian correlations equal to target Bernoulli correlations is generally incorrect after thresholding, and that pairwise tetrachoric calibration is exact only when the calibrated latent matrix is positive semidefinite. We therefore formulate the problem directly over the joint Bernoulli probability mass function. Given target means and pairwise correlations, we impose normalization, nonnegativity, mean constraints, and pairwise cross-moment constraints as a linear program over the $2^N$ atomic probabilities. The resulting PMF formulation either returns an exact law matching the requested first and second moments or certifies infeasibility. A convex-hull characterization further shows that every feasible target admits a law supported on at most $1+N+\binom{N}{2}$ states, while every infeasible target admits a separating quadratic certificate. We then develop a truncated-moment completion scheme that fits a reduced cross-moment table and generates samples by sequential conditioning, together with a sparse-support working-set refinement that can reduce memory usage on structured instances, although the worst-case complexity remains exponential. Together, these constructions provide an exact PMF-based framework for feasibility and simulation at moderate dimension and structured alternatives when the full atomic representation is impractical, while clarifying the limits of Gaussian-threshold constructions.
arXiv:2506.06570v4 Announce Type: replace
Abstract: As robotic systems become increasingly integrated into real-world environments, ranging from autonomous vehicles to household assistants, they inevitably encounter diverse and unstructured scenarios that lead to failures. While such failures pose safety and reliability challenges, they also provide rich perceptual data for improving system robustness. However, manually analyzing large-scale failure datasets is impractical and does not scale. In this work, we introduce the problem of unsupervised discovery of failure taxonomies from large volumes of raw failure logs, aiming to obtain semantically coherent and actionable failure modes directly from perceptual trajectories. Our approach first infers structured failure explanations from multimodal inputs using vision language reasoning, then clusters them in the resulting semantic reasoning space, discovering recurring failure modes rather than isolated episode-level descriptions. We evaluate our method across robotic manipulation, indoor navigation, and autonomous driving domains, demonstrating that the discovered taxonomies are consistent, interpretable, and useful in practice. In particular, we show that structured failure taxonomies guide targeted data collection for offline policy refinement and enhance runtime failure monitoring systems. Website: https://mllm-failure-clustering.github.io/
arXiv:2607.17288v1 Announce Type: new
Abstract: High quality temporal graph benchmarks with rich semantics and ground-truth anomaly labels are essential for training graph neural networks, yet remain scarce due to privacy constraints and annotation costs. We present SAGA (Synthetic Agentic Graph Architecture), a system for generating large-scale, semantically rich temporal graphs via a four-phase pipeline. Our Skeleton-First, Semantics-Second architecture decouples structure from semantics: (S) an O(1)-per-edge skeleton generator produces power-law graphs; (A) a dispatcher partitions causally ordered time blocks for parallel execution; (G) LLM agents inject domain semantics using RAG-based rule bases across four domains; and (A) a state alignment engine resolves conflicts via temporal replay, yielding anomaly labels as natural byproducts. Unlike structural generators (e.g., LDBC SNB, Kronecker/R-MAT) or purely LLM-based approaches, SAGA achieves structural realism, semantic richness, and automatic anomaly labeling in a unified framework. On a single H100 GPU with vLLM batching, SAGA generates 500,000 temporal edges with controlled anomalies in under 90 minutes, scaling to 100,000 nodes while maintaining clustering coefficients above 0.99. The system supports real-time pipeline visualization, interactive multi-domain tuning (Finance/AML, Network/IDS, Cyber/APT, Transportation), and a CLI for large-scale GPU-based experiments.
arXiv:2607.17063v1 Announce Type: new
Abstract: The rapid advancement of large language models (LLMs) has led practitioners to increasingly rely on them for answering questions about hardware description languages (HDLs). Because HDL is ultimately synthesized into physical hardware, an imprecise or redundant answer can propagate into timing violations or non-synthesizable logic that surface only late in the design flow, making the quality of HDL answers especially consequential. However, the quality of LLM-generated responses, particularly in comparison with answers provided by human experts, remains unclear. To investigate this question, we collect 6,246 HDL Q&A posts with accepted answers from Stack Overflow and curate them into a dataset, organized into a taxonomy of four main categories (Conceptual, Debugging, Generation, and Optimization) and ten subcategories. Using this dataset, we design a user study conducted with 19 HDL engineers with one to three years of experience. Our findings reveal a pervasive over answering tendency: LLMs supply correct content but bury it under redundant alternatives (65.7%) and verbose padding (69.1%), while nearly half of answers (49.0%) fail to fully align with expert answers yet participants still preferred LLM responses for readability (58.3%). Motivated by these findings, we propose a multi-agent framework for improving LLM-based HDL question answering. We evaluate answer quality using an LLM-as-Judge and two structural metrics: the number of core answers, which reflects redundancy since LLMs often provide multiple alternative solutions, and the length of non-core content, which reflects verbosity. Evaluated on the four mainstream LLMs, our framework increases the average core-answer quality score from 3.71 to 4.67 (+0.96) and the non-core content quality from 3.72 to 4.23 (+0.51), on a five-point scale.
arXiv:2607.17668v1 Announce Type: new
Abstract: Unsupervised Graph Domain Adaptation (UGDA) aims to facilitate knowledge transfer from a labeled source graph to an unlabeled target graph by mitigating cross-domain distribution shifts. Existing methods primarily focus on node-level feature alignment in latent spaces, relying on the implicit assumption that all source nodes contribute positively to the alignment. However, this assumption often fails because a node's semantic information is intrinsically coupled with its topological graph structure. Due to structural shifts, source nodes with severe structural deviations (e.g., structural outliers) lack semantic counterparts in the target graph, and forcing alignment on them introduces severe noise and causes negative transfer. To bridge this gap, we argue that selective source node utilization is superior to full-graph training, thereby shifting the research paradigm from feature-level alignment to data-level refinement. To this end, we propose Source Node Influence Pruning (SNIP), a novel model-agnostic, data-centric refinement framework. Specifically, SNIP quantifies the structural discrepancy between individual source nodes and the target domain by integrating multiple centrality measures, assigning each source node an influence score. A rank-based normalization mechanism is further employed to eliminate scale variations across different measures, allowing SNIP to effectively identify and filter out structurally incompatible nodes with low influence scores. As a plug-and-play method, SNIP constructs a refined "sub-source" graph that is inherently more beneficial for subsequent alignment. Comprehensive experiments across eight transfer scenarios on five real-world datasets demonstrate that SNIP consistently outperforms competitive baselines and significantly enhances adaptation performance, validating the superiority of selective node utilization over full-graph training.
arXiv:2607.16682v1 Announce Type: new
Abstract: The widespread adoption of high-level deep learning libraries, while accelerating model development, has increasingly abstracted away the internal mechanics of neural networks, creating a gap between practical usage and fundamental understanding. To address this, the paper presents a self-contained neural network framework implemented entirely from scratch -- without relying on automatic differentiation or pre-built deep learning modules. The implementation encompasses all essential components, including multi-layer architectures, diverse activation functions, regularization techniques, and state-of-the-art optimizers. Beyond serving as a pedagogical instrument that demystifies forward/backward propagation, gradient dynamics, and optimization landscapes, the framework demonstrates robust performance when applied to a multi-class classification task, successfully validating its correctness, numerical stability, and generalization across varied configurations. The extensible design and clean modularity further position it as a reliable baseline for educational purposes and future research exploration.
arXiv:2607.16889v1 Announce Type: new
Abstract: The neighbourhood complexity $nc(G,k)$ of a graph $G$ is a quantity measuring, for a graph $G$ and an integer $k$, the maximum possible number (over all vertex subsets $S$ of size $k$) $|\{N[v]\cap S, v\in V(G)\}|$ of $S$-neighbourhoods in $G$. This notion is important in structural graph theory and algorithm design (especially in parameterized complexity, in particular model checking and kernelization).
While generally $nc(G,k)\leq 2^k$ and this bound can be achieved, it is known that sparse graphs and structured dense graphs have linear neighbourhood complexity, that is, $nc(G,k)\in O(k)$ for any such graph $G$. However, for many graph classes, the best possible constants are not known. We focus on graphs of bounded treewidth and pathwidth, showing that (when $k\geq w+1$) (i) if $G$ has treewidth $w\geq 2$, then $nc(G,k)\leq (k-w+1)2^{w}+w$, and (ii) if $G$ has pathwidth $w\geq 2$, then $nc(G,k)\leq (k-w+2)2^{w-1}+2k-w-2$. Moreover, we provide constructions that reach these bounds, whenever $w\geq 2$ and $k\geq 2w+1$ ($k\geq 2w-1$ for pathwidth).
Interestingly, in contrast, we also have the tight bound $nc(G,k)\leq \frac{7}{3}k$, for graphs $G$ with pathwidth 1 or treewidth 1.
arXiv:2607.17292v1 Announce Type: new
Abstract: This project investigates the Phoa principle in synthetic domain theory (SDT), and provides a generalisation to the transfinite cases. The Phoa principle plays a pivotal role in SDT by illustrating how the paths give the information order on the interval type and other algebraic structures in SDT. The project defines the dual simplices and spines and introduces the concept of sobriomorphisms, which contributes to a new interpretation of the Phoa principle and its generalisations. Finally, the project proposes a hypothetical completeness theorem that may unify the Segal completeness and the chain completeness in SDT based on investigations on the Phoa principle in the project. The project also includes axiomatisation of the interval type in Cubical Agda and the formalised proof for the main theorems.
arXiv:2607.17673v1 Announce Type: new
Abstract: Contrastive learning is increasingly moving toward settings with three or more modalities instead of image-text pairs. Yet, extending models from pairwise to higher-order multimodal alignment can introduce optimization and representation challenges. We identify encoder Jacobian conditioning as a key factor in trimodal contrastive learning: poorly conditioned encoders exhibit collapsing or amplified singular-value spectra, leading to exploding Jacobian condition numbers and degraded multimodal alignment. We introduce geometry-preserving encoders (GPEs) by directly conditioning the Jacobian through regularization and demonstrating that simple modifications like LeakyReLU activations and residual paths recover these geometric benefits. Across a synthetic benchmark and four real-world datasets including missing modalities, improving Jacobian conditioning boosts retrieval and linear probe performance across multiple contrastive objectives, whereas expressive objectives yield little benefit in linear probes. More broadly, our results show that multimodal contrastive learning depends not only on objective expressivity, but also on the geometric and optimization properties of the underlying encoders.
arXiv:2506.07691v2 Announce Type: replace
Abstract: Sparse Autoencoders (SAEs) are a cornerstone of mechanistic interpretability. Existing training methods inherit the Block Training paradigm from LLM pre-training, which introduces destructive gradient noise in instruct models due to attention leakage from unrelated contexts. Using GSNR analysis, we theoretically characterize this issue and propose Finetuning-aligned Sequential Training (FAST), a sequential training paradigm specifically designed for instruct models. FAST aligns SAE training with the data distribution and activation patterns of instruct models, substantially improving both reconstruction fidelity and feature interpretability. Experimental results show that FAST achieves higher GSNR, a significantly lower log-scaled MSE of 0.6468 compared to the baseline's 5.1985, and a near-zero Delta Loss (-0.51\% to 0.37\%). Moreover, on Llama-3.2-3B-it, FAST produces 21.1\% high-quality features, substantially outperforming baseline methods that achieve 7.0\% and 10.2\%. We further find that intervening on special token activations through SAEs can improve generation quality, revealing new opportunities for fine-grained control. Our codes are available as open source at https://github.com/Geaming2002/FAST.