arXiv:2607.12094v1 Announce Type: new
Abstract: Reliable detection of out-of-distribution (OOD) samples is crucial for the safe deployment of machine learning models. Neural networks often produce overconfident predictions for inputs that deviate from their training data, leading to significant degradation in performance. While many OOD detection methods focus on the final output layer, they neglect the rich hierarchical information present in intermediate network layers. This paper introduces a novel approach that leverages sparse autoencoders (SAEs) to learn interpretable features from these intermediate activations. We find that in-distribution (ID) and OOD data activate distinct sets of these sparse features. We propose a new OOD score derived from the cosine similarity between the sparse feature activations of a test sample and the mean activations of ID classes. Our post-hoc detection method not only achieves state-of-the-art performance on standard OOD detection benchmarks, but yields interpretable insights into how distribution shift affects learned representations.
Science Journals
arXiv:2607.12104v1 Announce Type: new
Abstract: Machine learning models for system diagnostics rely on kernel execution traces to capture fine-grained system behavior, but collecting production traces in industrial systems is costly due to runtime overhead, storage demands, and privacy constraints. We present TraceSynth, a diffusion-based framework for generating synthetic kernel traces that augment limited real data for downstream ML tasks. TraceSynth models traces as multi-channel sequences (event types, timestamps, CPU affinity, thread identifiers, and process metadata) using a Transformer-based denoising diffusion process with constraint-guided repair to enforce system invariants. Across six benchmarks, results show strong workload dependence. For deterministic, compute-heavy workloads (scimark2), synthetic augmentation achieves 87.2% F1-Macro at context length L=4096, only 2.6 percentage points below real-only baselines. Context length is the dominant quality factor, with L=4096 yielding a +104% relative improvement over L=256, while constraint-guided repair improves synthetic data quality by up to 4.3%. Ablation studies show that lightweight 2-channel models retain 97-99% of the performance of full 6-channel models at roughly half the computational cost. TraceSynth supports cost-effective augmentation of kernel execution traces in production observability pipelines and helps identify when synthetic data can substitute for limited real traces.
arXiv:2607.12372v1 Announce Type: new
Abstract: Multimodal semantic segmentation (MSS) is essential for robust perception in complex environments, yet its potential remains largely untapped because of the prohibitive cost of human annotations. While unsupervised semantic segmentation (USS) has achieved strong results on a single RGB modality, its naive extension to multimodal data is often hindered by fusion degradation. This occurs because, without explicit supervision, existing frameworks struggle to reconcile the heterogeneous structural patterns captured by different sensors and therefore fail to effectively exploit their complementary information. In this paper, we make the first attempt to address the novel problem of Unsupervised Multimodal Semantic Segmentation (UMSS), aiming to effectively exploit complementary sensor information in a fully label free setting. To this end, we propose UniM2 (Unified Multimodal), a novel framework built on DINOv3 that transforms conventional fusion methods into consistent performance gains. Our key idea is to learn a unified latent space driven by Cross Modal Correspondence Synergy (CMCS) to extract intrinsic shared semantic cues, bypassing the need for label guided adaptive fusion. To mitigate inherent intermodal conflicts, we introduce a Cross Modal Harmonizer (CMH) that designates RGB as a stable reference, effectively suppressing inconsistent relational supervision while guiding the model to exploit complementary structural features. Extensive experimental results on NYU Depth v2 and MFNet show that UniM2 improves mIoU by 6.4% and 9.8%, respectively, demonstrating clear advantages over existing frameworks for UMSS.
arXiv:2607.12385v1 Announce Type: new
Abstract: A significant challenge in agentic AI is prospective memory: the ability to execute an intention at a specific future cue or state while other activities are ongoing. We introduce PM-Bench, a text-based benchmark for measuring prospective memory capabilities in modern LLM agents. Inspired by the Virtual Week paradigm from cognitive science, PM-Bench evaluates how well LLM agents maintain user intentions, execute delayed intentions, and monitor latent environment changes. Over the course of a simulated seven-day week, agents must continue an ongoing activity while deciding whether any deferred task is due. We compare eight state-of-the-art LLMs on PM-Bench under eight different agent configurations. PM-Bench proves challenging across all settings: the best method, a GPT-5.4 agent, reaches only 65.1\% F1 score under our evaluation. Furthermore, no single strategy for improving prospective memory dominates across models. We release PM-Bench as a controlled testbed for diagnosing these failures and developing training or inference-time interventions that support reliable prospective behavior.
arXiv:2607.05668v2 Announce Type: replace
Abstract: Multiphysics simulation with lattice Boltzmann methods (LBM) requires a scheme hand-derived for each partial differential equation (PDE), a labor-intensive, error-prone bottleneck. We recognize our recently proposed class of LBM schemes as a discrete-kinetic relaxation approximation of conservation laws and generalize its hand derivation to an automated one for systems of hyperbolic, parabolic, and mixed-type conservation laws.
The derivation splits into three steps: First, the PDE system is equivalently rearranged into a first-order cascade of conservation laws: every spatial derivative in flux or source becomes an auxiliary variable, recursively for higher derivatives, so all fluxes are algebraic and updates stay local. Second, the augmented system is approximated by a discrete-velocity kinetic relaxation model with linear, constant-coefficient transport: all nonlinearity resides in a local equilibrium embedding the flux exactly in its first moment, trading the low-Mach truncation for an a priori checkable sub-characteristic wave-speed bound. Third, the relaxation system is discretized by a standard LBM, yielding collide-and-stream algorithms running unchanged on existing solvers.
A symbolic compiler using a domain-specific language encapsulates these steps: unlike existing LBM code generators, which start from the discrete scheme, it automatically derives equilibrium, gradient-tracking cascade, and grid scaling from the declared PDE alone. We exercise it across twelve PDE systems, including compressible Navier--Stokes--Fourier flow, resistive magnetohydrodynamics, and nonlinear elasticity. Manufactured-solution verification confirms convergence at or near second order in double precision, retained in single precision by a reference- and equilibrium-shifted formulation. Targeting OpenLB, the generated GPU kernels reach up to 96% of the memory-bandwidth roofline.
arXiv:2607.11318v2 Announce Type: replace-cross
Abstract: Wegner conjectured that every finite family $\mathcal R$ of axis-parallel rectangles satisfies $\tau(\mathcal R)\le 2\nu(\mathcal R)-1$, where $\nu$ is the packing number and $\tau$ is the piercing number. Ajwani, Gajjala, Raman, and Ray recently disproved this by constructing a triangle-free counterexample on $2196\cdot 8^9$ rectangles and, using a computer-assisted package-and-port recursion, obtained a standard LP gap of $17891/8064$ for Maximum Independent Set of Rectangles. We give a simpler and hand-checkable counterexample with $64$ rectangles. It is built from an eight-rectangle gadget whose independent sets inject into four ordered slots; we then use four horizontal and four vertical copies of this gadget to form a triangle-free family with $\nu=16$ and $\tau\ge 32$. We use the same horizontal-vertical step to define recursive families of rectangles $P_r$ with $\nu(P_r)=4^{2^r}$. For the standard clique, equivalently point, relaxation we obtain a finite gap $73/32$ at $P_3$, improving the previous benchmark of $17891/8064$. We then construct recursive fractional solutions and matching piercing sets showing $\lim_r \alpha^*(P_r)/\nu(P_r)=\lim_r \tau(P_r)/\nu(P_r)=5/2$. Finally, by disjoint union with isolated rectangles, we show that every rational $t\in[1,5/2)$ occurs as a standard LP gap and also as a packing-piercing ratio for suitable rectangle families.
arXiv:2607.12185v1 Announce Type: new
Abstract: Embodied accounts of semantic memory highlight the role of sensorimotor systems in acquiring and storing knowledge. Congenitally blind populations offer a critical test bed for these assumptions, providing an opportunity to assess whether conceptual grounding requires visual experience. In this study, we assessed semantic memory navigation differences between blind and sighted individuals using a property listing task with concrete and abstract concepts. We computed semantic entropy, an embedding-based natural language processing metric that captures the predictability of retrieval. Generalized linear mixed models revealed distinct navigation patterns across groups: while sighted individuals showed higher entropy for abstract than concrete concepts, blind participants did not. Instead, blind individuals exhibited higher entropy for visually salient concrete concepts (e.g., penguin). These results underscore the role of visual experience in the organization and dynamic navigation of semantic memory.
arXiv:2607.12050v1 Announce Type: new
Abstract: Multi-robot teams operating in confined or cluttered environments must adapt both their formation geometry and group topology to navigate through complex obstacles. This adaptation requires two complementary behaviors: deformation, where the team continuously reshapes its geometry while remaining connected, and reconfiguration, where robots split into subgroups or merge back into a single formation. Existing methods often model these behaviors independently, connect them through handcrafted rules, or lack explicit geometric criteria for determining when each behavior should be invoked. However, challenging environments may require online changes in formation shape, connectivity, and effective team composition, making decoupled or rule-based approaches prone to suboptimal trajectories and deadlock. We propose EFLUX, a geometry-grounded LLM agentic framework for automatic and elastic multi-robot formation navigation. EFLUX extracts a structured scene representation and uses an LLM to reason jointly over both deformation actions, such as scaling and shearing, and reconfiguration actions, such as splitting and merging. These strategies are then translated into executable per-robot waypoints through a closed-loop generation, verification, and correction pipeline. Simulation and hardware experiments show that EFLUX enables safe, continuous, and elastic formation navigation in constrained environments, reducing deadlock and navigation failures compared with baselines while maintaining coherent multi-robot coordination.
arXiv:2607.12187v1 Announce Type: new
Abstract: Many recently proposed Calderbank-Shor-Steane (CSS) quantum low-density parity-check (QLDPC) codes have sparse decoding graphs, enabling syndrome-based belief propagation (BP) decoding at low complexity. Their construction, however, often results in properties that impair BP performance, such as short cycles and degeneracy. In this work, we propose a general framework for introducing auxiliary variable nodes (AVNs) and auxiliary check nodes (ACNs) into the decoding graph of CSS codes, compatible with the standard stabilizer measurement framework. This provides an additional degree of freedom in the design of the decoding graph itself and can be used to tackle the aforementioned shortcomings. We show that recently proposed techniques, 4-cycle removal and subcode ensemble decoding, can be interpreted as instances of this framework. For 4-cycle removal, we find that the gains depend strongly on the BP iteration count and check-node message scaling. Building on this framework, we further propose a graph-derived subcode ensemble decoder and demonstrate under circuit-level noise that it substantially reduces the per-round logical error rate compared with BP on the corresponding 4-cycle-free decoding graph.
arXiv:2607.12954v1 Announce Type: new
Abstract: Engineering use of AI forecasting models requires not only high nominal accuracy but also predictable behavior under uncertain inputs. In photovoltaic (PV) forecasting, this requirement is especially challenging because numerical weather prediction (NWP) errors are temporally correlated, state dependent, and physically coupled across variables. Existing evaluations, however, often rely on perfect forecast assumptions or simplistic perturbations that do not reflect these characteristics. This study presents a physically constrained robustness evaluation framework based on simulation, using virtual PV power as a controlled response variable to isolate the propagation of input uncertainty from confounders at the plant level. Six representative machine learning and deep sequence models, including PatchTST, GRU, N-HITS, and LightGBM, are evaluated under dynamic NWP perturbations with heteroscedasticity modulated by clear-sky conditions and Erbs reconstruction that preserves radiation consistency. The results show that sequence models provide stronger noise filtering and temporal resilience than a strong tabular baseline under medium to high disturbance regimes. SHapley Additive exPlanations (SHAP) and Integrated Gradients (IG) further support a feature reallocation tendency at the case level, in which predictive reliance shifts from corrupted future forecasts toward more stable historical observations and deterministic physical priors. A Pareto analysis of accuracy under clean conditions, robustness, and computational latency then translates these findings into engineering implications for robustness assessment and model selection under forecast uncertainty.
arXiv:2607.12368v1 Announce Type: new
Abstract: Numerical simulation of the Vlasov-Poisson system faces fundamental challenges due to phase-space filamentation. Standard spectral methods rely on artificial filtering to suppress errors, which inadvertently degrades physical structures and accuracy over time. This paper proposes a dynamically adaptive Fourier spectral method combined with high-order time splitting to overcome these limitations. To prevent filamentation-induced aliasing, we dynamically expand the wavenumber domain via adaptive zero-padding whenever spectral tail monitoring detects emerging fine-scale structures. Acting as an exact trigonometric interpolation, it inherently avoids the artificial smoothing of traditional filters, preserving mass and effectively maintaining the $L^2$-norm within the dynamically resolved scales. Numerical experiments on Landau damping and various instability problems validate the robustness of our scheme.
arXiv:2607.11954v1 Announce Type: new
Abstract: Clinical code prediction maps unstructured discharge summaries to ICD-10-CM leaf codes in a large, sparse, and deeply hierarchical label space. Most systems treat the task as flat multi-label classification, scoring codes independently and providing limited training signal for rare labels. We propose a graph-constrained traversal policy that formulates ICD prediction as a finite-horizon decision process over a pruned code hierarchy. A single language model descends the graph level by level, selecting valid child nodes until billable leaf codes are reached. This converts extreme multi-label prediction into sparse, hierarchy-aware subset decisions while guaranteeing structurally valid outputs.
On MIMIC-IV discharge summaries, our best supervised policy, SFT-1+, achieves 0.709 micro-F1 on a curated 50-code subset and 0.527 micro-F1 on the full 15,761-code space, outperforming flat baselines including CAML, LAAT, and PLM-ICD. In the full setting, SFT-1+ improves over the strongest flat baseline by 0.044 micro-F1 and 0.157 macro-F1, suggesting that graph-constrained decomposition mitigates the rare-code bottleneck. A controlled factorial study evaluates architecture, training algorithm, and data budget. Across both scales, one shared policy matches a three-specialist cascade while avoiding its context-window overflow on 28-32% of full-space test notes. Increasing supervised trajectory data is the only intervention that consistently improves performance, while GRPO reinforcement learning provides no benefit over supervised continuation with matched data. These results show that simple graph-constrained policy learning can outperform more complex flat, cascaded, and reinforcement-learning alternatives for extreme clinical code prediction.
arXiv:2607.12221v1 Announce Type: new
Abstract: Industry is leaning into generative artificial intelligence (GenAI), and higher education is under pressure to prepare graduates for a GenAI-augmented workforce. Yet, there is still no clear structure for defining AI readiness across disciplines, programs, courses, and assignments. Current approaches often rely on broad institutional policies or individual course-level decisions, which can also create mixed messages for students, fragmented expectations across programs, and limited visibility for university leaders. In this paper, we argue that higher education needs a more coherent way to connect institutional priorities to curriculum-level action. We propose Program-Level AI Learning Outcomes (PLAI-LOs) as a framework for defining what students graduating from a program should know and be able to do with, without, and about GenAI in a given discipline. The PLAI-LOs framework complements existing program-level learning outcomes and supports alignment across institutional priorities, program-level AI learning outcomes, course-level learning outcomes, and assignment-level objectives. We illustrate the framework with examples from computing and music and show how PLAI-LOs can be implemented through artifact-level GenAI policies, helping programs decide where GenAI should be taught and used, and when students should be expected to work without GenAI. We offer PLAI-LOs as a concrete, measurable, and adaptable path for moving higher education from scattered GenAI rules toward a strategy with clear, learning-centered alignment.
arXiv:2607.11963v1 Announce Type: new
Abstract: The early detection of Chronic Kidney Disease using machine learning has attracted significant interest in healthcare-related computer science. Despite rapid advancements in this field, many reported studies remain inconsistent and potentially misleading. A significant drawback is the lack of organized evaluation regarding methodological concerns. Key issues include data leakage, limited access to temporal patient records and inconsistency in reported clinical indicators. This research offers a systematic literature review of existing CKD prediction studies using interpretable machine learning techniques, where nineteen relevant studies were selected via systematic searches across major academic databases. To assess methodological reliability, this study introduces a structured taxonomy of information leakage and a quantitative leakage scoring framework to systematically evaluate reliability across CKD prediction studies. The analysis reveals a strong relationship between leakage and inflated performance. Here, High leakage-studies report an average accuracy of 95.48%, compared to 80.2% for leakage-free studies, reflecting an increase of approximately 15.28%. Furthermore, a cross-study feature stability analysis shows that only a small subset of predictors is consistently reproducible, with over 80% lacking reliability. Overall, the findings suggest that many reported performance improvements stem from methodological limitations rather than true predictive capability.
arXiv:2607.11974v1 Announce Type: new
Abstract: Most neural partial differential equation (PDE) surrogates learn how fields evolve after a grid has already been chosen. However, before any operator is applied, the grid has already determined how modeling capacity is allocated across space, resolution, and spectral bandwidth. We argue that this hidden design choice should itself be learnable, leading to a question different from standard operator learning: can a surrogate learn where resolution should exist before predicting field evolution? We formulate adaptive discretization as a physics-constrained conditional generation problem over valid mesh displacements. The success of diffusion models in PDE field prediction suggests their potential for learning adaptive discretizations under similar structured constraints. This leads to a two-stage diffusion framework: Stage 1 learns an r-adaptive displacement mesh conditioned on the observed dynamics, while Stage 2 predicts the solution evolution from the mesh-informed representation. The mesh generator is regularized by physics-aware proxy channels, geometric validity constraints, and local spectral concentration so that adaptation remains physically interpretable and numerically legal. Across five PDE regimes, the results show that diffusion-based learned discretization is competitive with adaptive-mesh and reduced-order baselines, with particularly strong gains in regimes where fixed or handcrafted allocation is insufficient. The main conclusion is not that there exists a universal optimal mesh rule, but that discretization should be learned in a regime-dependent manner: different spatial and spectral structures favor different allocation behaviors. This reframes adaptive meshing for neural PDE solvers from a solver-specific heuristic into a generative representation-learning problem.
arXiv:2607.12222v1 Announce Type: new
Abstract: AI data centers are emerging as a dominant new load class with their power dynamics fundamentally from conventional industrial loads. Inside a training job, the bulk-synchronous-parallel algorithm moves each node through compute, sync, and checkpoint steps, which swings power between full load and near idle within seconds. Across the whole facility, jobs arrive, take blocks of nodes for hours to days, then leave, so the number of busy nodes changes daily, weekly, and yearly. This slower shift drives facility-wide swings and the peak demand that sets the size of the grid link. A model that looks only at within-job behavior, and treats the facility as a fixed set of busy nodes, smooths out these swings and misses the true peak-to-average ratio. This paper develops a hierarchical semi-Markov Data-Center (HSM-DC) load model that couples two layers across two timescales. A job-scheduling layer creates jobs through a non-homogeneous compound-Poisson process shaped by daily, weekly, and seasonal patterns, gives each job a heavy-tailed node count and length, and places jobs on a fixed pool of nodes on a first-come basis. A within-job layer moves each busy node through a five-state semi-Markov chain for the BSP steps, with state-based Ornstein-Uhlenbeck noise. Facility power comes from this changing node count and the per-node power, set to match measured node data and the facility's straight-line power-versus-load curve. Configured to the reference facility at the same scale, the model matches mean power, its spread, and the peak-to-average ratio across load levels, with fit scores of 0.9997, 0.92, and 0.82. It also matches the share of queued jobs to within one point at high load. Facility-wide swings and peak demand come from how jobs arrive and get scheduled, so grid planning must model that process, not just scale up a single node's power curve.
arXiv:2607.12233v1 Announce Type: new
Abstract: Large language model (LLM) trading agents show promising performance in equity markets, yet remain narrowly focused on US equities with little evidence from live deployment. We present Fin-Analyst, a hybrid agent for FinMMEval 2026 Task 3: an eight-specialist LLM pipeline over news, SEC filings, fundamentals, analyst forecasts, technical indicators, and social sentiment, aggregated by a Meta-Agent for Tesla (TSLA), and a lightweight rule based three-signal vote for Bitcoin (BTC). On the final official leaderboard (accessed 2026-07-05), Fin-Analyst ranks first of all agents on TSLA with a +13.51% return, +28.33 points over Buy-and-Hold (Sharpe 4.10, 88% win rate), while the BTC vote ends flat yet well above a sharply falling baseline. Relative to the interim performance, the asset ranking reversed, indicating that short live windows yield volatility-sensitive rankings. Ablation identifies event-driven 8-K disclosures as the most influential TSLA signal. Error analysis shows that the memoryless agents repeat wrong calls for days at a time, and that the fixed-threshold BTC rules lost money by trading on noise in a sideways market while the LLM pipeline gained under similar conditions, motivating a memory-aware, LLM-based successor for both assets.
arXiv:2607.12172v1 Announce Type: new
Abstract: Decentralized gradient descent (DGD) is widely used for solving distributed optimization problems over networks of agents. While its convergence properties are well understood, less is known about the communication and computation resources required to attain a prescribed accuracy. In this paper, we study DGD from a resource-aware perspective and characterize the communication-computation budget required to attain a target error level. We develop a bottleneck-centric framework in which different factors dominate the optimization dynamics at different error scales. Specifically, we identify operating regimes governed by initialization, objective heterogeneity and network connectivity, gradient noise, and communication noise. To capture these effects, we introduce two fundamental quantities: the gradient-Diversity-to-Network-connectivity Ratio (DNR) and the Gradient-to-Communication-noise Ratio (GCR). We show that these quantities determine the sequence of bottlenecks encountered during optimization and the corresponding budget-optimal operating strategy. Using a multi-stage analysis, we derive optimal stepsize selections and explicit budget-complexity bounds that quantify the budget resources required to attain a prescribed accuracy. The resulting expressions reveal how the overall budget decomposes into contributions associated with successive bottlenecks and provide insight into the fundamental tradeoffs among objective heterogeneity, network connectivity, gradient noise, and communication noise.
arXiv:2607.12182v1 Announce Type: new
Abstract: Annihilation photons are quantum entangled in their polarization, a property that is not accessible in state-of-the-art clinical Positron Emission Tomography (PET). This roadmap describes the current status of research in the emerging field of quantum entanglement applications involving these photons for medical diagnosis. It outlines the underlying physics phenomena and the development of detector systems that can serve as a foundation for future Quantum Entanglement PET (QE-PET) scanners. These scanners will be capable of utilizing the entanglement between annihilation photons by measuring their Compton scattering on electrons. This roadmap comprises up-to-date experimental results on the study of quantum entanglement and the decoherence of annihilation photons, alongside the current theoretical understanding of these phenomena. The methods and detector technologies described herein are being developed (i) in order to enhance standard PET imaging by suppressing random coincidences in the reconstruction of annihilation site density distributions, (ii) in order to establish the degree of quantum entanglement as a diagnostic biomarker for tissue pathology and oxygenation, and (iii) in order to elaborate a method for pH imaging. Whether entanglement-based imaging and pH mapping can be successfully translated into clinical practice remains an open question and a subject of exciting ongoing research. This roadmap serves as an invitation to the scientific community to join this burgeoning field.
arXiv:2607.12237v1 Announce Type: new
Abstract: E-textiles and wearable sensing technologies enable flexible, customizable interfaces for human-computer interaction, with capacitive sensing offering precise touch and pressure detection. While machine knitting provides scalable, mechanically tunable structures ideal for such sensors, few studies develop or characterize insulated conductive yarns engineered for knitting's complex structural geometry and high flexure strain. In this work, we present a yarn dip-coating process, driven by an adjusted dip-coating fluid dynamics model, that enables scalable, machine knittable fabrication of capacitive tactile pressure sensing arrays. We establish optimal dip-coating parameters and concentrations of thermoplastic polyurethane (TPU) dissolved in dimethylformamide (DMF) to create knitting-optimized coatings (~630 um thickness). These fabricated yarns are shown to maintain electromechanical characteristics with minimal deviation after knitting and washing, thus allowing the creation of knitted pressure sensors through multi-layered structures. This process demonstrates that machine knitting with insulated yarns is a viable and reliable manufacturing approach to integrate sensing functionality into wearable textiles.
arXiv:2607.12240v1 Announce Type: new
Abstract: Structural optimization problems often involve a large number of decision variables and highly non-convex feasible regions, making convergence to the true Pareto front extremely challenging. Even when convergence is achievable, it typically requires thousands of function evaluations, resulting in significant computational cost. This highlights the need for efficient and robust optimization algorithms for real-world engineering applications. In this study, we introduce a novel constrained multi-objective evolutionary algorithm, termed DPCME. The algorithm employs two interacting populations that exchange information, enabling effective global exploration and reducing the risk of convergence to local optima. To further enhance performance, a recent repair-based constraint-handling technique is incorporated, and alternative repair approaches are proposed and systematically evaluated. The proposed algorithm is tested on three engineering problems: the 72-bar truss, the 120-bar truss, and a chemical tanker structure, each involving hundreds of nonlinear failure constraints. Its performance is evaluated against state-of-the-art constrained multi-objective optimization algorithms from the latest PlatEMO package. A total of 43 algorithms are initially tested, from which the 12 best-performing methods are selected for detailed comparison. The results demonstrate that DPCME achieves superior or competitive convergence and diversity across all test cases, and that the inclusion of repair-based constraint handling further improves its performance.
arXiv:2607.12245v1 Announce Type: new
Abstract: Few-shot industrial defect detection remains difficult for standard supervised detectors, which achieve poor performance on boundary-dominated industrial defects. This paper proposes rough path signature-guided geometry augmentation (RPS-GA), a geometry-aware approach in which Canny edge contours are treated as ordered planar paths whose truncated second-order signature responses, especially the antisymmetric L\'evy-area term, are aggregated into a spatial map that highlights boundary-related structure through two fusion operators, SIG-AUG and SGAA. The approach is evaluated on NEU-DET and PCB-Defect under a few-shot protocol with 5, 10, 20, or 50 labeled images per class, using an unmodified YOLOv8n detector throughout. Compared with the baseline, RPS-GA delivers large gains when supervision is limited, although the margin shrinks as more labels become available. On NEU-DET, SIG-AUG raises 10-shot mAP@0.5 from 0.341 to 0.583, whereas on PCB-Defect, SGAA improves 10-shot mAP@0.5 from 0.086 to 0.299 and yields usable detection at 5-shot where the baseline fails entirely. These trends are confirmed by multi-seed evaluation across independent random partitions. Overall, the results indicate that second-order path-signature geometry offers a practical way to strengthen few-shot industrial defect detection without meta-learning or detector redesign.
arXiv:2607.12257v1 Announce Type: new
Abstract: On-device research agents search a corpus, read sources, and write a cited brief on a personal laptop. Whether their citations are faithful, and at what cost, is unmeasured for a deployable small model. This study fixes one 4B generator on a 24 GB laptop and asks what makes its citations faithful. It separates two quantities usually reported as one number. Cited claim faithfulness asks whether the cited source supports the claim. Trustworthy coverage asks whether the agent also cites the right sources. The study crosses how much of each source the generator sees, 400 against 1500 characters, with the quality of the sources supplied, gold papers against retrieved papers. Two levers fall out, and they act on different outcomes. Exposure sets faithfulness. More of each source lifts faithfulness from 0.45 to 0.58 on retrieved sources and from 0.37 to 0.58 on gold sources, and the two settings converge, so faithfulness is bound by exposure, not by whether the source is correct. The exposure lift is robust to a second, independent judge; the exact convergence is tight under the primary judge and only approximate under the second. Retrieval sets coverage. Trustworthy coverage stays near 0.22 on retrieved sources at any exposure, because recall is held near 0.40, so exposure cannot fix which sources are cited. The extra exposure costs about 235 output tokens. The practical recipe is to raise per source exposure first, cheaply, and then treat retrieval recall as the only remaining lever.
arXiv:2607.12615v1 Announce Type: cross
Abstract: We study the limits of third-degree price discrimination when the production cost is Bayesian and private to the seller, generalizing the seminal work of Bergemann, Brooks and Morris (2015). The rough setup is the following: A monopoly seller sets different prices for buyers in different "segments" of the market so as to maximize seller surplus. Different ways in which the aggregate market is decomposed into segments lead to different welfare outcomes, i.e., (seller surplus, buyer surplus) pairs. When the production cost is Bayesian, the region of achievable welfare outcomes can exhibit complex shapes beyond the clean characterization by Bergemann, Brooks and Morris for the case with a fixed cost. We show that with a Bayesian cost, this region coincides with a proper projection of a polytope defined by a polynomial number of linear constraints, the essential ones of which correspond to flow conservation in a "discounted" flow network. As a result, we give a polynomial-time algorithm that computes optimal market segmentations in terms of any linear combination of the seller surplus and the buyer surplus. En route, we establish the following structural property: Any market can be written as a convex combination of "extremal markets" in a way preserving the seller surplus and the buyer surplus. These extremal markets are piecewise equal-surplus with respect to different possible costs, generalizing a similar notion introduced by Bergemann, Brooks and Morris when the cost is fixed.
arXiv:2511.18089v2 Announce Type: replace
Abstract: Multimodal survival analysis aims to improve cancer prognosis using heterogeneous biomedical data, such as histopathology images and genomic profiles. A common strategy is to align representations across modalities so that shared signals can be captured. However, strong cross-modal alignment can also remove modality-specific evidence that is critical for survival prediction. In this paper, we revisit multimodal survival learning from a simple observation: effective models should first discover shared patterns across modalities, and then preserve modality-specific signals. This motivates a representation learning principle that we refer to as Together Then Apart. Based on this idea, we propose TTA, a framework that balances cross-modal alignment and representation distinctiveness. TTA first performs prototype-based alignment to capture shared survival-related structures between modalities. It then encourages modality-specific distinctiveness through an anchor-guided contrastive objective. To further account for modality imbalance and noisy correspondences, we model cross-modal interactions using unbalanced optimal transport. We evaluate the proposed approach on multiple TCGA cancer cohorts with paired histopathology and genomic data. TTA consistently improves survival prediction over recent multimodal survival models. Moreover, the learned prototype structures reveal interpretable cross-modal patterns associated with clinical outcomes.