Papers

Efficient acquisition of correlated continuous-time signals is critical in applications such as distributed sensor networks and array processing. However, existing correlation models often fail to capture the shared structure of physical signals measured in close proximity. In this paper, we model each signal in an ensemble as the sum of a hidden common lowpass component and a signal-specific innovation highpass component with disjoint spectral supports, where the common bandwidth is unknown. We first derive theoretical identifiability conditions guaranteeing unique decomposition and reconstruction from subsampled observations. Further, we propose a practical recovery algorithm and a joint reconstruction framework that leverages parametric structured dictionaries parameterized by the unknown common bandwidth. Numerical experiments on an ensemble of four signals demonstrate exact reconstruction with a 25% aggregate sampling rate reduction under identifiability conditions, and robust recovery with up to a 74% rate reduction when all channels are sampled below the Nyquist rate, significantly reducing data acquisition and hardware overhead.

Post-training quantization (PTQ) is a practical approach to reducing the memory and computational footprint of large language models (LLMs) without retraining. GPTQ-based methods have become the de facto standard, yet they suffer from two complementary limitations. Methods with local, layer-wise objectives lack global supervision; while methods with global objectives fix their Hessian estimates at the start and ignore first-order gradients, so their guidance grows stale as quantization proceeds. This paper presents G$^2$PTQ, a unified PTQ framework with Generalized Gradient Compensation that integrates both first- and second-order information under a globally supervised, block-wise optimization objective. By refreshing gradient and Hessian estimates before quantizing each Transformer block, G$^2$PTQ avoids the staleness of prior global methods. Furthermore, to stabilize the exact first-order compensation, we introduce a trust-region scaling mechanism that dynamically bounds the gradient step to prevent exploding weight updates. Finally, we derive efficient implementations for block-wise Hessian approximation and exact gradient compensation. Experimental results on various model families and bit-widths demonstrate that G$^2$PTQ enables better alignment with the full-precision model, outperforming state-of-the-art baselines. Code is available at: https://github.com/G2PTQ/G2PTQ.

Quadruped robots typically rely on depth cameras and LiDAR sensors to map their local environment. However, these sensors have limited close-range coverage, are relatively expensive, and consume significant power. This study investigates whether distributed Time-of-Flight (ToF) sensors can serve as a low-cost alternative to depth cameras for near-field terrain mapping for locomotion and local navigation. We designed a distributed ToF sensing architecture for the ANYbotics ANYmal quadruped, assessed its environment reconstruction accuracy, and benchmarked it against depth cameras for terrain mapping and obstacle avoidance. Distributing these sensors around the robot can also avoid the blind spots of traditional sensors. Our results show that, despite their low resolution and higher measurement noise, distributed ToF sensors can support reliable perceptual locomotion with centimeter-level local mapping accuracy. The proposed sensing strategy provides sufficient geometric information for near-field obstacle avoidance and footstep planning, at substantially lower cost, energy consumption, and system complexity than depth cameras.

Large language models can produce fluent prose from short prompts, but direct prompt-to-text interaction gives writers limited access to the assumptions that shape a long narrative. We present Incipit, an implemented research prototype that introduces an explicit planning layer between a writer's intent and generated prose. This layer is grounded in literary axioms, defined as curated and reusable propositions about human experience and narrative craft. The prototype connects a knowledge base of 1455 axioms and 472 typed relationships with a five-round direction dialogue, a retrieval-and-selection pipeline, and a three-level blueprint covering creative premises, story beats and character arcs, and chapter outlines. Writers can inspect and edit these structures before using them as context for scene generation. Additional modules support real-event abstraction and five-dimensional diagnostic feedback. We describe the system design rationale, data flow, implementation boundaries, and a worked design example. As no controlled user study or independently rated output study has yet been completed, we do not claim that the system improves literary quality. Instead, we outline a future preregistered evaluation designed to distinguish the contribution of axiom grounding from that of hierarchical planning. The paper contributes a concrete architecture for using literary knowledge as an inspectable coordination object in human-AI writing.

Anti-altermagnets (AAMs) are a recently identified class of layered magnetic materials where opposite spin-splitting in adjacent layers creates a globally spin-degenerate band structure, rendering conventional spectroscopic detection highly difficult. In this paper, we theoretically demonstrate an all-electrical method to manipulate and probe this hidden magnetic order using a dual-gated transport junction. By applying a perpendicular displacement field, we break the spatial inversion symmetry of the lattice, explicitly lifting the global spin degeneracy. We attach ferromagnetic (FM) leads on either side of the AAM. Using quantum transport calculations, we show that this gate-induced spin-splitting manifests as a strongly asymmetric tunnel magnetoresistance (TMR) as a function of the lead magnetization. We identify specific crystallographic orientations where the TMR retains its symmetry despite the fully split bands, a direct consequence of exact momentum-space compensation. We further reveal that Rashba spin-orbit coupling guarantees robust, highly directional transport asymmetries even in the absence of an explicit chemical potential mismatch between the layers. Our findings establish a clear, electrically tunable framework for exploiting AAMs in next-generation spintronic architectures.

Open-vocabulary 3D maps enable robots to reason about previously unknown environments using natural language. However, existing systems typically segment every incoming image, associate detections with persistent 3D segments, and frequently perform costly Vision-Language (VL) inference. We present TRACKGRAPH, an online open-vocabulary system that maintains short-term 2D mask identity directly in the image stream before fusing segments into 3D. FastSAM masks and CLIP features are computed at sparse keyframes, while dense DINOv3 features are used to propagate masks at a high rate in between. The resulting tracked masks are fused into a class-agnostic 3D segment layer within a hierarchical scene graph, with 3D association handling tracking interruptions and long-term revisits. Compact multi-view CLIP embeddings enable open-vocabulary retrieval. Across Replica, ScanNet++, and HM3D, TRACKGRAPH achieves competitive open-vocabulary segmentation and retrieval against state-of-the-art mapping methods, including the highest synonym frequency on Replica (0.50). On the same NVIDIA A100, it is 1.7x faster and uses 3.3x less GPU memory than ViT-H OVI-MAP. Real-world quadruped deployments demonstrate onboard scene graph construction and object search at 7.5Hz, while recorded drone data is used to test the method under aerial viewpoints.

Attackers can compromise multiple systems with a single vulnerability, while defenders need to fix all security weaknesses in their systems. This asymmetry puts defenders at a disadvantage. Security vulnerabilities are found at an alarming rate, and patching vulnerabilities is costly and time-consuming; thus, vulnerability prioritization is a must and a time-critical challenge. Many prioritization metrics, such as the CVSS, EPSS, KEV, and SSVC, are currently used, each with different pros and cons, such as openness, degree of automation, time-criticality, coverage, and need for expert input. In this work, we propose the GitHub Popularity Metric (GPM), a fully open, publicly computable prioritization metric based on the popularity of exploits in GitHub repositories. We use GitHub features such as the number of stars, forks, and related unique users to create a metric that indicates the popularity of CVEs across different time frames, both relative to the current time and historically. We compare the proposed metric with various existing vulnerability prioritization metrics and known exploited vulnerabilities and demonstrate that it provides tangible insights for defenders, identifying unique CVEs and inconsistencies in existing methods. The GPM was integrated into the EPSS version 5. \noindentNote. A demonstration based on this work has been accepted to the demo track of the ACM Conference on Computer and Communications Security (CCS) 2026.

Machine-learning (ML)-based malware detectors are commonly trained using labels obtained from antivirus (AV) engines and aggregation services (e.g., VirusTotal). This practice assumes AV-generated labels provide reliable supervision. However, small byte-level modifications can substantially alter AV verdicts while leaving the representations perceived by downstream ML detectors largely unchanged, producing label-feature inconsistencies that can contaminate training datasets and create poisoning opportunities for ML-based malware detection. We present Bi-Iocane, a black-box poisoning framework that exploits the reliance of malware-labeling pipelines on AV-generated labels. Bi-Iocane identifies AV-sensitive bytes and modifies them to induce label changes. It rewrites such bytes in malware to obtain benign labels (evasion-oriented poisoning) and injects malware-associated byte patterns into benign software to obtain malicious labels (defamation-oriented poisoning). These poisoned samples and their lightly modified variants corrupt training data and cause selected targets to be misclassified. We evaluate Bi-Iocane with 13 AV engines simulating AV aggregation services and eight ML detectors. For 30 malware and 30 benign clean targets, Bi-Iocane combines AV-specific manipulations to generate malware-to-benign and benign-to-malware poisoned samples whose all tested AV-based labels are flipped. After these poisoned samples and variants are used for downstream training, the resulting ML models misclassify 92.08% of the original clean targets on average with only a 0.06% poisoning budget per target. Meanwhile, the poisoned models largely preserve clean-set performance, and six evaluated poisoning defenses show only limited mitigation. VirusTotal evaluation further confirms practical defamation risk and reveals potential evasion risk in real-world AV-to-ML labeling supply chains.

Open rerankers trained for general web retrieval transfer imperfectly to e-commerce, where ranking decisions depend not only on topical relevance but also on user preferences, product constraints, and comparative product fit. These preference signals are difficult to supervise at scale: real search traffic provides authentic queries and candidates but no clean pairwise labels. We present ZooWork-ShopRanker, a family of e-commerce rerankers (0.6B, 4B, and 8B) aligned to judge-labeled shopping preference. Training pairs are labeled by a panel of reasoning large language models (LLMs) from different families acting as a preference oracle, with position-debiased judgments and agreement tiers, and the rerankers are trained on these labels. The aligned 8B flagship then serves as a distillation teacher for the efficient 4B and 0.6B models, which are fit to its scores and sharpened on judged pairs. To measure progress, we introduce ShopRank-Bench, a contamination-limited benchmark of ~10,000 private-traffic preference pairs in both text formats, tiered by how many judge families committed to each label. ZooWork-ShopRanker-8B and -4B significantly outperform the strongest open reranker baseline, every model significantly beats its own un-aligned base, and ZooWork-ShopRanker-0.6B beats its size peer; the gains hold in both formats and extend to common MTEB benchmarks. We release the models and the dual-format ShopRank-Bench to facilitate further research.

CeRh2As2 is a heavy-fermion superconductor that exhibits distinct low- and high-field superconducting phases, widely interpreted as reflecting a field-induced change in the parity of the order parameter. However, the superconductivity develops in close proximity to magnetic order whose microscopic origin and nature remain unclear. Moreover, the low-lying crystal-electric-field-split doublets of the Ce 4f electrons appear to realize an unusual "quasi-quartet" structure. This has motivated proposals that quadrupolar degrees of freedom play an important role in the low-temperature physics. Here we develop a microscopic heavy-fermion description of CeRh2As2 that incorporates both the quasi-quartet structure and the nonsymmorphic symmetry. Within a multiorbital periodic Anderson model, we include the lowest-lying $\Gamma_7$ and $\Gamma_6$ doublets of the Ce 4f electrons, and allow for nonlocal hybridization with the conduction d electrons. Single occupancy of the 4f states is enforced within the rotationally invariant slave-boson formalism. Adopting a mean-field treatment, we first show that the effective low-energy Hamiltonian is accurately captured by a single-orbital tight-binding model, justifying the minimal descriptions widely employed in the literature. Allowing for a magnetic ground state, we find that our model displays metallic $\mathbf{Q}=(\pi,\pi)$ antiferromagnetic order over a wide parameter regime. Ferromagnetic solutions also occur, but in a smaller region of parameter space and with greater sensitivity to carrier concentration and the form of the f-d hybridization. Quadrupolar moments appear only within magnetically ordered phases and are generally small, becoming substantial only when the $\Gamma_7$ and $\Gamma_6$ doublets are nearly degenerate. Our results provide a microscopic basis for understanding the itinerant heavy-fermion physics and ordered phases of CeRh2As2.

In this paper, we investigate the periodic orbits around a black hole (BH) in Einstein-Maxwell-scalar (EMS) theory and the corresponding gravitational waveforms. First, we study the test particle dynamics around the BH by using the Lagrangian formalism. Using the corresponding conditions, we explore the effect of the spacetime parameters on the marginally bound orbit (MBO) and innermost stable circular orbit (ISCO). The results indicate that MBO and ISCO quantities decrease with the rise of the $\beta$ parameter and BH charge $Q$, and vice versa for the $\gamma$ parameter. Subsequently, we plot the different zoom-whirl orbits around the BH in EMS theory by finding the test particle's corresponding energy and angular momentum. Finally, the gravitational waveforms of the extreme mass ratio inspirals, which contain a supermassive BH in EMS theory and a stellar mass object, are demonstrated. It is shown that the zoom-whirl orbits shrink as the BH charge increases, and, as a result, gravitational waveforms shift towards shorter times.

We establish an exact additive formula expressing the Lelong number of a plurisubharmonic function in terms of the limiting difference between its spherical and solid means. Building on this identity, we derive several equivalent criteria for the vanishing of the Lelong number, formulated in terms of fixed-scale increments and the asymptotic agreement of radial maxima, spherical and solid means, and radial regularizations. We also give a pointwise $L^1$ mean-oscillation formulation of the VMO characterization for plurisubharmonic functions with zero Lelong number, together with a streamlined proof of the corresponding global result of Biard and Wu. Finally, we establish a nonlinear characterization in terms of the relative extremal depth and the Bedford-Taylor capacity of deep sublevel sets.

Processing-using-DRAM (PUD) architectures exploit the analog operational properties of DRAM to perform bulk bitwise Boolean and arithmetic operations inside memory arrays by organizing data in a vertical layout, where operand bits are stacked along DRAM columns. However, modern computing systems natively employ a horizontal data layout that preserves the cache line abstraction, leverages spatial locality in row buffers, and enables high memory throughput. This fundamental mismatch forces existing PUD architectures to frequently perform data layout transformations between horizontal and vertical formats, incurring significant performance, energy, and system integration overheads. Our goal is to eliminate data transposition overheads in PUD systems at low cost. To this end, we propose PipeDRAM, a PUD architecture that eliminates the need for runtime data layout transformation, enabling PUD operations directly over horizontally laid-out data. PipeDRAM's key ideas are to (i) deterministically reorganize bits inside each memory request to enable a PUD-friendly data placement within a DRAM array in a horizontal data layout, and (ii) employ a pipeline-based execution model that overlaps bit-dependent and bit-independent in-DRAM operations to exploit bit-level parallelism across the memory array. We compare PipeDRAM to different computing platforms. PipeDRAM provides (i) 11.8x, 11.8x, and 80.4x higher performance and (ii) 25.4x, 3.0x, and 38.0x lower energy consumption than three state-of-the-art PUD systems. PipeDRAM incurs low area cost on top of a DRAM chip (1.86%) and CPU die (0.05%). To enable further research on PUD systems, we open-source PipeDRAM at https://github.com/CMU-SAFARI/PipeDRAM.

Passive image provenance asks whether pixels alone can reveal where an image came from: a human, an aggregate AI class, or a particular generator. This becomes a robustness problem once a source image can be edited before the verifier sees it. We study the problem as source–target verification under adversarial distribution shift. Our first result gives the exact best-case limit for any image-only verifier: the largest robust target-acceptance gap equals the minimum total-variation distance between the target distribution and the set of attacked source distributions. This quantity depends on the source, target, and edit class, not on the verifier architecture. Our second result explains why deployed public verifiers can fail before this statistical limit is reached. If the verifier can be emulated on the attack region to error $\varepsilon$, then a surrogate black-box attack reaches target acceptance within $2\varepsilon$ plus optimization error of the white-box optimum; score-revealing logistic and softmax heads over public features are identifiable, and approximate score access gives stable recovery bounds. A finite-state experiment checks the minimax identity where both sides are computable. On same-prompt real/diffusion benchmarks, the evaluated public CLIP verifiers fail under targeted pixel attacks, while a ResNet-18 victim exhibits partial fake-to-real transfer. Binary feedback with abstention reduces measured attack success, but positive empirical gap upper bounds do not establish robustness. These results motivate separate evaluation of the source–target statistical ceiling and the information released by a deployed verifier.

The linear representation hypothesis (LRH) has become a standard lens for measuring and intervening on semantic information through the internal representations of large language models (LLMs). A growing body of work has begun extending this perspective to vision-language-action (VLA) models, but the dynamical nature of embodied interaction introduces an additional challenge. Unlike semantic attributes commonly studied in LLMs, such as gender or language, a physical quantity of interest (QoI) in a VLA evolves jointly with the system dynamics: the representation influences the actions selected by the policy, which alter the physical state and, in turn, the next representation. In this paper, we develop a theoretical, signature-based formulation of the LRH for VLA that unifies representations and policies. On the representation side, we establish the existence of representations from which the future evolution of a QoI under a candidate action trajectory can be recovered via linear probing. On the policy side, we introduce a signature generalized linear model for stochastic action chunks. This structure yields a monotonic change in the expected future QoI along linear paths in natural parameter space, enabling linear steering. We first validate this structure in a controlled oracle setting, then examine whether the same probing and steering mechanisms emerge in a pretrained VLA.

Machine learning (ML) emulators provide a fast and cost-effective method to simulate climate change scenarios after being trained on Earth System Models projections. However, the black-box nature of those data-driven approaches limit the usability and trustworthiness of their outputs and in particular their use as causal attribution tools. Here, we develop a hierarchical causal representation learning framework applied to sea surface temperature fields from a state-of-the-art global climate model. As a key advance over previous work, our framework explicitly models both atmospheric dynamical interactions arising from internal climate variability and forced responses due to changes in atmospheric greenhouse gas and aerosol concentrations. When trained on future climate change scenarios, our method accurately predicts the long-term global mean and regional temperature evolution and shows physically realistic responses to perturbations in greenhouse gas and aerosol concentrations when evaluated on unseen scenarios. Our results underline the potential of causal representation learning frameworks for advancing climate model emulation.

Considering a cylindrically symmetric spacetime, a novel exact $\kappa$-deformed Schwarzschild-AdS black string solution is derived in General Relativity in the presence of a negative cosmological constant. The geometric properties of the resulting solution are systematically investigated by computing the Ricci and Kretschmann curvature scalars. The analysis reveals the existence of a curvature singularity at the origin, which remains enclosed by an event horizon provided that the $\kappa$-deformation parameter satisfies a specific constraint. Subsequently, the influence of various model parameters on the horizon structure and configuration is examined in detail. The thermodynamic properties of these black strings are thoroughly evaluated; in this regard, the Hawking temperature, entropy, mass, isobaric heat capacity, and Gibbs free energy are self-consistently extracted. The effects of the parameters, particularly the $\kappa$-deformation parameter, on these thermodynamic quantities are comprehensively explored. Furthermore, it is demonstrated that these state variables fundamentally satisfy the extended first law of black hole thermodynamics and yield a modified Smarr relation. Finally, the temporal profile of the Hawking radiation is probed via the dimensionless sparsity parameter, through which the impacts of the underlying parameters on the quantum emission rate are systematically analyzed.

We present FLIP, a final-layer inference-time probe for testing whether a logit-facing intervention site in an open-weight vision-language model (VLM) supports structured, task-linked computation rather than generic perturbation. Behavioral change under internal intervention is otherwise mechanistically ambiguous: it may reflect improved use of visual evidence, generic output instability, or outright degradation. FLIP applies elementwise flooring to the final normalized hidden state before logit computation, leaving parameters, prompts, and decoding unchanged. On a controlled detection/counting probe, sweeping intervention strength reveals three regions: negligible change, a bounded interior regime in which detection recall at IoU 0.50 ($R_{50}$) improves while tolerant counting error ($\mathcal{E}_{\mathrm{count}}$) falls, and over-suppression. We formalize a four-criterion probe-and-sweep protocol for disciplining the interpretation of intervention effects: regime structure, grounding-proxy alignment, feature-coherence dependence, and failure to reproduce the same positive regime on a performance-based negative control. The post-normalization state passed to the output head is the logit-facing instantiation of this test; under a non-targeted flooring sweep it satisfies the full protocol. Raw decoder-layer interventions, including the last-block output before final normalization, and the singleton-pair left/right control fail to reproduce the Final-site signature, while same-site operators and multiple VLMs replicate it. FLIP is therefore a validation step for intervention-based mechanistic interpretability, not a steering method. Repo at https://github.com/earl-juanico/flip-qwen3vl/

We give a deterministic $O(m)$-time algorithm for exact modular subset sum over every modulus $m$. The input gives distinct residues and their multiplicities. The algorithm reports all reachable residues and answers one target query. It returns a witness when the target is reachable. The algorithm uses $O(m)$ auxiliary words on an arithmetic word-RAM. This improves Potępa's deterministic $O(m\log m\,\alpha(m))$-time bound under the same input convention. The running time matches the cost of returning a length-$m$ Boolean array. We represent reachable residues as runs along cycles of repeated addition. Newly reached residues pay for scans of partial runs. Packed cycle flags make rebuilding take linear total time at the original modulus. A theorem on subset sums of distinct units limits the number of adaptive boundary batches to $O(m^{3/4})$. Radix sorting their $O(m)$ total keys also takes $O(m)$ time.

The feedback-based algorithm for quantum optimization (FALQON), which is an extension of the quantum approximate optimization algorithm (QAOA), is a promising approach for finding approximate solutions to combinatorial optimization problems. The FALQON avoids the optimization of variational parameters by using information derived from iterative measurements. To solve a constrained combinatorial optimization problem, it is sometimes more efficient to use an XY mixer rather than the conventional X mixer. We applied the FALQON with the XY mixer to graph coloring and investigated the expectation value of the problem Hamiltonian, success probability of graph coloring, and fair sampling for several small-scale problems. The FALQON with the XY mixer exhibited better performance than that with the X mixer.

The Teukolsky Master Equation (TME) is a powerful technique used to study perturbations of a spacetime with gravitational, electromagnetic, and neutrino components. The well-known Kerr metric is a suitable candidate for this formalism because it is a Petrov type D metric. However, the Kerr metric represents the exterior spacetime of a perfect sphere and only black holes are described using this metric. For another astrophysical object, such as fast-spinning neutron stars, the Kerr metric is not valid due to the rotation induced deformations to the gravitational source. The main goal of the present work is to show a TME for a Kerr-like metric (KL) with a mass quadrupole moment parameter $q$. This new parameter will represent deformations from a perfect sphere; therefore, KL is a better candidate to study perturbations surrounding spinning stars.

Purpose: Accurate 6 DoF pose estimation of surgical tools is critical for automa- tion, robotic proprioception, and safe interaction with the tissue operated on. Kinematics-based approaches suffer from accumulated errors due to the cable- driven nature of robotic arms, while vision-based methods often rely on external markers or trackers. Although more recent vision-based advances have been pro- posed, these two-stage pose estimation methods often lack real-time robustness due to accumulated errors and computational overhead. Methods: We propose a novel end-to-end trainable model, PICO. Our model employs a multi-task learning architecture to predict segmentation and depth maps, alongside regression of translation and rotation parameters. We define two proxy tasks that enforce geometric consistency in both 2D and 3D spaces, improving accuracy and robustness. For this, we propose a projection loss, and a point-to-point loss. Results: We evaluate our method on the SurgRIPE dataset, benchmarking its performance against state-of-the-art approaches using standard 6DoF pose esti- mation metrics. Our results demonstrate consistently strong performance across all four subsets, specifically in rotation, ranking second even under occlusion. It also demonstrates comparable translational performance, remaining competitive, especially in occluded cases. Conclusion: PICO demonstrates the effectiveness of multi-task learning and geometry-aware proxy tasks for robust and reliable surgical tool pose estimation, especially in occluded scenarios, highlighting potential for future applications.

Real-world image super-resolution (SR) requires recovering perceptually realistic high-resolution images from complex low-resolution observations while preserving faithful content. Diffusion-based SR benefits from strong generative priors but incurs substantial computational overhead, whereas feed-forward CNN and Transformer SR models are efficient yet often struggle to recover realistic high-frequency details. This motivates a natural question: can diffusion priors be transferred to existing diffusion-free SR networks without introducing diffusion components at inference time? To this end, we propose PhoenixSR, a generative heterogeneous distillation framework that transfers diffusion priors to independently designed feed-forward SR networks through score-based distribution matching. Rather than aligning heterogeneous features or imitating sampled diffusion outputs, PhoenixSR uses the pretrained diffusion model as distribution-level supervision, while paired SR supervision preserves reconstruction fidelity. To make distribution matching effective for fidelity-sensitive SR, we introduce Heterogeneous Distribution Adaptation, which adapts the target score to the SR domain, improves tracking of the evolving student distribution, and anchors training with paired supervision. We further employ Directional Reliability Weighting, a lightweight residual-consistency-based reweighting strategy that reduces unstable distributional guidance. All diffusion-related components are removed after training, leaving the original student architecture and inference cost unchanged. Experiments on three SR benchmarks and six feed-forward backbones, including SwinIR, HAT, Real-ESRGAN, and SeeMoRe, show consistent perceptual improvements with largely preserved reconstruction fidelity.

Self-supervised monocular depth estimation (MDE) enables depth prediction from monocular images without requiring ground-truth supervision, making it attractive for large-scale and real-world applications. Despite steady improvements in accuracy, most existing methods remain difficult to interpret, as depth is inferred from RGB representations that obscure the impact of individual perceptual image components. This lack of transparency limits systematic analysis of failure cases and reduces confidence in safety-critical settings. This paper presents a self-supervised framework for perceptually interpretable monocular depth estimation (PIMDE), designed to associate depth predictions with distinct perceptual components of the input image. Rather than operating directly on RGB inputs, the proposed method decomposes each image into a set of perceptual feature maps (PFMs), each encoding a specific visual cue. Distinct depth estimation branches process these PFMs independently to produce depth estimates (PIDEs), which are subsequently combined through an explicit fusion strategy. This formulation allows us to examine directly the contribution of each perceptual cue to the final depth prediction. Experiments conducted on the KITTI benchmark dataset demonstrate that PIMDE achieves performance comparable to established self-supervised MDE methods while providing additional insight into how different perceptual cues influence depth estimation. These results indicate that perceptual decomposition can support interpretability without sacrificing depth estimation accuracy.

As large language models increasingly mediate information access, factually accurate and independent answers are critical. However, these models can exhibit sycophancy by aligning their responses with users' stated beliefs even when those beliefs are incorrect, potentially presenting misinformation as independently verified and reinforcing users' confidence in false claims. Prior work leaves unresolved whether introducing user beliefs causes correct responses to become incorrect or uncertain, or causes uncertain responses to become belief-aligned incorrect answers. It also remains unclear whether anti-sycophancy interventions preserve or restore factual accuracy or merely shift responses toward uncertainty. We analyze factual sycophancy in Chinese-language information seeking using yes/no fact-checking questions. Our analysis covers 364,941 responses from three frontier Chinese-based LLMs (DeepSeek, Qwen, and Doubao) to 12,165 factual questions derived from real-world Chinese search queries. We evaluate the models with and without reasoning across baseline, belief-conditioned, and anti-sycophancy prompting, tracing matched shifts among correct, incorrect, and uncertain responses. Under incorrect user beliefs, we distinguish belief-aligned errors from losses of factual confidence, in which initially correct answers become uncertain. Patterns vary across models and reasoning settings: reasoning is not a consistent safeguard, and anti-sycophancy instructions can reduce incorrect agreement while increasing uncertainty. In Chinese-language factual question answering, avoiding agreement with false beliefs is therefore not equivalent to preserving factual accuracy, highlighting the value of transition-level evaluation. Such behavior may undermine the reliability of LLM-mediated information access by reinforcing misinformation or weakening users' confidence in factually correct answers.

Mutational signatures reveal cancer-driving processes with clinical relevance: MMR-deficient tumors respond better to immunotherapy, HRD tumors are sensitive to PARP inhibitors, and POLE-mutant tumors often have high mutation burdens influencing treatment response. Yet signature profiling remains limited by the cost, complexity, and turnaround time of whole-genome sequencing (WGS) or whole-exome sequencing. Histopathology is widely available and cost-effective, and H&E morphology has been linked to MSI, HRD, POLE-related processes, and driver mutations. Whether histology can recover the broader landscape of mutational signature exposures in a pan-cancer setting remains unknown. Here we introduce Hist2Sig, a deep learning framework predicting exposures to 30 COSMIC SBS signatures from H&E-stained slides. Trained on matched WGS and histology data from 7,063 TCGA patients across 29 cancer types, Hist2Sig was compared with a tumor type-only baseline to separate morphology-derived signal from tissue-of-origin priors. In internal cross-validation, it recovered relative signature compositions across most tumor types and achieved higher mean top-three overlap than the baseline in 19 of 29, with the largest gains in COAD, PRAD, GBM, and UCEC. It also recurrently recovered signatures of poorly characterized etiology, including SBS8, SBS12, SBS39, and SBS40a. In an external CPTAC cohort of 193 patients across five tumor types, Hist2Sig retained moderate concordance with observed compositions and outperformed the baseline in glioblastoma and pancreatic adenocarcinoma. These results support the feasibility of inferring mutational signature exposures from routine histology, while highlighting variation across tumor types. Hist2Sig provides a foundation for tumor-specific models that could complement sequencing in pre-sequencing triage or support decisions where WGS is unavailable.

Text-to-speech needs written text in spoken form, and speech recognition output needs the reverse. For Khmer, neither direction has a maintained open-source tool, and the script makes both harder: words are not separated by spaces, and number words occur inside ordinary words. We present Tha, a Khmer text normalization and inverse text normalization toolkit built from weighted finite-state transducers. It segments and classifies a whole line in one shortest-path search, and a second transducer rejects token boundaries inside a Khmer syllable. On Google's Khmer test suite, Tha agrees with the reference on all 274 cardinals up to one spelling variant, and on 2,906 real TTS prompts, 153 of the 158 sentences it rewrites are correct. Tha is open source under the Apache 2.0 license.

Recent audio deepfake detectors separate bona fide speech from synthetic speech, yet it remains unclear which stage of a text-to-speech system supplies the detection evidence. We address this with controlled resynthesis and detector adaptation in an F5-TTS-BigVGAN pipeline. Since vocoder reconstruction of a real mel can itself be separable from the source utterance, we fix the vocoder and trace the larger change in detector separation to acoustic generation. Adversarially fine-tuning the acoustic model, with no detector in its objective, raises EER against fixed detectors at comparable quality. However, adapting a detector only on the tuned model's VCTK outputs lowers its LibriSpeech EER from 19.42% to 7.46% and improves detection of unseen base F5-TTS outputs. These results suggest that acoustic-model updates can reduce the detection evidence available to fixed detectors, while detector adaptation keeps the updated outputs detectable in this pipeline.

Modern AI image generators are increasingly deployed as opaque APIs, where customers can query the deployed service, but cannot inspect model weights or architecture. This creates a practical challenge: a provider may pass governance certification with one generator and later silently switch to a cheaper and lower-quality one for deployment, compromising public trust or even safety in high-stakes domains. We study integrity auditing at deployment time and propose FARE (Forensic Acceptance Region Estimation). A certified generator is enrolled by training FARE on images sampled from that generator. After deployment, FARE can determine whether a generated image is consistent with the enrolled generator—using only that image. FARE's features are based on image generator-specific artifacts that have been proposed for forensic applications. FARE amplifies these features during training by finding hard samples that tighten the acceptance region and increase sensitivity to subtle changes in the certified generator. Across generator swaps, including substitutions with similar model versions and model variants, FARE is effective at detecting swaps, consistently outperforming existing baselines at strict operating points, and remains effective under the exact-model and decision-only attacks evaluated in this work.

Reliable online video question answering requires tracking state transitions while selectively abstaining when visual evidence is insufficient. Existing benchmarks focus on static recognition or long-range retrieval, rarely evaluating these coupled capabilities under evolving and incomplete evidence. We present STORM-Bench, comprising 5,736 questions across 630 compact, change-dense episodes spanning five egocentric domains (STORM-Real) and two controlled simulation subsets (STORM-Sim) at 1 FPS. Questions are stratified by a proxy for accumulated change intensity (Low, Medium, High) and query-time answerability (Known, Uncertain). To measure reliability, we introduce STORM-BR, a harmonic metric over joint answer-status correctness that exposes abstention failures masked by aggregate accuracy, alongside STORM-BR-ATTR for uncertainty attribution. Across 14 video LLMs, online accuracy peaks at 60.3% (mean 51.7%), whereas STORM-BR ranges from 5.7% to 35.6% (mean 18.8%), driven by pervasive overconfidence on uncertain queries. STORM-Bench shows that task accuracy masks these gaps in epistemic reliability and state tracking. Benchmark and code are available at https://github.com/siruzhong/STORM-Bench.