Accurate estimation of the body surface area (BSA) involved by a rash, such as psoriasis, is critical for assessing rash severity, selecting an initial treatment regimen, and following clinical treatment response. Attempts at segmentation of inflammatory skin disease such as psoriasis perform markedly worse on darker skin tones, potentially impeding equitable care. We assembled a psoriasis dataset sourced from six public atlases, annotated for Fitzpatrick skin type, and added detailed segmentation masks for every image. Reference models based on U-Net, ResU-Net, and SETR-small are trained without tone information. On the tuning split we sweep decision thresholds and select (i) global optima and (ii) per Fitzpatrick skin tone optima for Dice and binary IoU. Adapting Fitzpatrick specific thresholds lifted segmentation performance for the darkest subgroup (Fitz VI) by up to +31 % bIoU and +24 % Dice on UNet, with consistent, though smaller, gains in the same direction for ResU-Net (+25 % bIoU, +18 % Dice) and SETR-small (+17 % bIoU, +11 % Dice). Because Fitzpatrick skin tone classifiers trained on Fitzpatrick-17k now exceed 95 % accuracy, the cost of skin tone labeling required for this technique has fallen dramatically. Fitzpatrick thresholding is simple, model-agnostic, requires no architectural changes, no re-training, and is virtually cost free. We demonstrate the inclusion of Fitzpatrick thresholding as a potential future fairness baseline.

We adopt several renowned backdoor defense strategies in ANNs to SNNs and analyze the challenges blocking them from being as effective in SNNs. Based on those findings, we propose innovative designs to solve the identified challenges to defending against backdoor attacks in SNNs with neuromorphic data. We propose TMPBD, a novel data-free, unsupervised backdoor detection strategy based on the TMP's MM statistic, which reaches 100% attack label detection accuracy on models poisoned by various backdoor attacks without access to any data. We propose NDSBM, a novel unsupervised backdoor mitigation strategy based on clamping the weights of the connection, also known as neural dendrites in SNNs, between the first two convolution layers in each convolution block of the model. NDSBM is capable of lowering the ASR from 100% down to 8.44% on average against dynamic trigger attacks. In addition, we utilize the end-to-end backdoor defense pipeline for both proposed backdoor detection and mitigation strategies to further reduce the ASR under SOT A dynamic trigger attack to 2.81% on average while achieving higher CA. We comprehensively evaluate the proposed backdoor defense strategies against the existing defense methods adopted for ten repetitions with multiple attack types and variant datasets. We critically discuss the scalability and robustness of the proposed methods against imbalanced datasets and adaptive attackers and provide indicative solutions to false-positive, intrinsic backdoor, and all-to-all attack issues when additional information are available.

The performance of machine learning models depends heavily on training data. The scarcity of large-scale, well-annotated datasets poses significant challenges in creating robust models. To address this, synthetic data generated through simulations and generative models has emerged as a promising solution, enhancing dataset diversity and improving the performance, reliability, and resilience of models. However, evaluating the quality of this generated data requires an effective metric. This paper introduces the Synthetic Dataset Quality Metric (SDQM) to assess data quality for object detection tasks without requiring model training to converge. This metric enables more efficient generation and selection of synthetic datasets, addressing a key challenge in resource-constrained object detection tasks. In our experiments, SDQM demonstrated a strong correlation with the mean Average Precision (mAP) scores of YOLOv11, a leading object detection model, while previous metrics only exhibited moderate or weak correlations. Additionally, it provides actionable insights for improving dataset quality, minimizing the need for costly iterative training. This scalable and efficient metric sets a new standard for evaluating synthetic data.

While modern deep neural networks achieve impressive performance in vision tasks, they remain opaque in their decision processes, risking unwarranted trust, undetected biases and unexpected failures. We propose cluster paths, a post-hoc interpretability method that clusters activations at selected layers and represents each input as its sequence of cluster IDs. To assess these cluster paths, we introduce four metrics: path complexity (cognitive load), weighted-path purity (class alignment), decision-alignment faithfulness (predictive fidelity), and path agreement (stability under perturbations). In a spurious-cue CIFAR-10 experiment, cluster paths identify color-based shortcuts and collapse when the cue is removed. On a five-class CelebA hair-color task, they achieve 90% faithfulness and maintain 96% agreement under Gaussian noise without sacrificing accuracy. Scaling to a Vision Transformer pretrained on ImageNet, we extend cluster paths to concept paths derived from prompting a large language model on minimal path divergences. Finally, we show that cluster paths can serve as an effective out-of-distribution (OOD) detector, reliably flagging anomalous samples before the model generates over-confident predictions. Cluster paths uncover visual concepts, such as color palettes, textures, or object contexts, at multiple network depths, demonstrating that cluster paths scale to large vision models while generating concise and human-readable explanations.

Generative AI leaderboards are central to evaluating model capabilities, but remain vulnerable to manipulation. Among key adversarial objectives is rank manipulation, where an attacker must first deanonymize the models behind displayed outputs -- a threat previously demonstrated and explored for large language models (LLMs). We show that this problem can be even more severe for text-to-image leaderboards, where deanonymization is markedly easier. Using over 150,000 generated images from 280 prompts and 19 diverse models spanning multiple organizations, architectures, and sizes, we demonstrate that simple real-time classification in CLIP embedding space identifies the generating model with high accuracy, even without prompt control or historical data. We further introduce a prompt-level separability metric and identify prompts that enable near-perfect deanonymization. Our results indicate that rank manipulation in text-to-image leaderboards is easier than previously recognized, underscoring the need for stronger defenses.

Neural models such as YOLO and HuBERT can be used to detect local properties such as objects ("car") and emotions ("angry") in individual frames of videos and audio clips respectively. The likelihood of these detections is indicated by scores in [0, 1]. Lifting these scores to temporal properties over sequences can be useful for several downstream applications such as query matching (e.g., "does the speaker eventually sound happy in this audio clip?"), and ranked retrieval (e.g., "retrieve top 5 videos with a 10 second scene where a car is detected until a pedestrian is detected"). In this work, we formalize this problem of assigning Scores for TempOral Properties (STOPs) over sequences, given potentially noisy score predictors for local properties. We then propose a scoring function called LogSTOP that can efficiently compute these scores for temporal properties represented in Linear Temporal Logic. Empirically, LogSTOP, with YOLO and HuBERT, outperforms Large Vision / Audio Language Models and other Temporal Logic-based baselines by at least 16% on query matching with temporal properties over objects-in-videos and emotions-in-speech respectively. Similarly, on ranked retrieval with temporal properties over objects and actions in videos, LogSTOP with Grounding DINO and SlowR50 reports at least a 19% and 16% increase in mean average precision and recall over zero-shot text-to-video retrieval baselines respectively.

Online communities rely on a mix of platform policies and community-authored rules to define acceptable behavior and maintain order. However, these rules vary widely across communities, evolve over time, and are enforced inconsistently, posing challenges for transparency, governance, and automation. In this paper, we model the relationship between rules and their enforcement at scale, introducing ModQ, a novel question-answering framework for rule-sensitive content moderation. Unlike prior classification or generation-based approaches, ModQ conditions on the full set of community rules at inference time and identifies which rule best applies to a given comment. We implement two model variants - extractive and multiple-choice QA - and train them on large-scale datasets from Reddit and Lemmy, the latter of which we construct from publicly available moderation logs and rule descriptions. Both models outperform state-of-the-art baselines in identifying moderation-relevant rule violations, while remaining lightweight and interpretable. Notably, ModQ models generalize effectively to unseen communities and rules, supporting low-resource moderation settings and dynamic governance environments.

We consider an important problem in scientific discovery, namely identifying sparse governing equations for nonlinear dynamical systems.

Previous work has focused mainly on bounding either the expected loss of a predictor or the probability that an individual prediction will incur a loss value in a specified range.