Large Language Models (LLMs) are susceptible to adversarial attacks such as jailbreaking, which can elicit harmful or unsafe behaviors.This vulnerability is exacerbated in multilingual settings, where multilingual safetyaligned data is often limited.Thus, developing a guardrail capable of detecting and filtering unsafe content across diverse languages is critical for deploying LLMs in real-world applications.In this work, we introduce a multilingual guardrail with reasoning for prompt classification.Our method consists of: (1) synthetic multilingual data generation incorporating culturally and linguistically nuanced variants, (2) supervised fine-tuning, and (3) a curriculum-based Group Relative Policy Optimization (GRPO) framework that further improves performance.Experimental results demonstrate that our multilingual guardrail, Mr-Guard, consistently outperforms recent baselines across both in-domain and out-of-domain languages by more than 15%.We also evaluate MrGuard's robustness to multilingual variations, such as code-switching and low-resource language distractors in the prompt, and demonstrate that it preserves safety judgments under these challenging conditions.The multilingual reasoning capability of our guardrail enables it to generate explanations, which are particularly useful for understanding languagespecific risks and ambiguities in multilingual content moderation.

Uncertainty in the predictions of learning-enabled components hinders their deployment in safety-critical cyber-physical systems (CPS). A shift from the training distribution of a learning-enabled component (LEC) is one source of uncertainty in the LEC’s predictions. Detection of this shift or out-of-distribution (OOD) detection on individual datapoints has therefore gained attention recently. But in many applications, inputs to CPS form a temporal sequence. Existing techniques for OOD detection in time-series data for CPS either do not exploit temporal relationships in the sequence or do not provide any guarantees on detection. We propose using deviation from the in-distribution temporal equivariance as the non-conformity measure in conformal anomaly detection framework for OOD detection in time-series data for CPS. Computing independent predictions from multiple conformal detectors based on the proposed measure and combining these predictions by Fisher’s method leads to the proposed detector CODiT with bounded false alarms. CODiT performs OOD detection on fixed-length windows of consecutive time-series datapoints by using Fisher value of the input window. We further propose performing OOD detection on real-time time-series traces of variable lengths with bounded false alarms. This can be done by using CODiT to compute Fisher values of the sliding windows in the input trace and combining these values by a merging function. Merging functions such as Harmonic Mean, Arithmetic Mean, Geometric Mean, Bonferroni Method, and so on, can be used to combine Fisher values of the sliding windows in the input trace, and the combined value can be used for OOD detection on the trace with bounded false alarm rate guarantees. We illustrate the efficacy of CODiT by achieving state-of-the-art results in two case studies for OOD detection on fixed-length windows. The first one is on an autonomous driving system with perception (or vision) LEC. The second case study is on a medical CPS for walking pattern or GAIT analysis where physiological (non-vision) data is collected with force-sensitive resistors attached to the subject’s body. For OOD detection on variable length traces, we consider the same case studies on the autonomous driving system and medical CPS for GAIT analysis. We report our results with four merging functions on the Fisher values computed by CODiT on the sliding windows of the input trace. We also compare the false alarm rate guarantees by these four merging functions in the autonomous driving system case study. Code, data, and trained models are available at https://github.com/kaustubhsridhar/time-series-OOD .

Incorporating learning based components in the current state-of-the-art cyber-physical systems (CPS) has been a challenge due to the brittleness of the underlying deep neural networks. On the bright side, if executed correctly with safety guarantees, this has the ability to revolutionize domains like autonomous systems, medicine, and other safety-critical domains. This is because it would allow system designers to use high-dimensional outputs from sensors like camera and LiDAR. The trepidation in deploying systems with vision and LiDAR components comes from incidents of catastrophic failures in the real world. Recent reports of self-driving cars running into difficult to handle scenarios is ingrained in the software components which handle such sensor inputs. The ability to handle such high-dimensional signals is due to the explosion of algorithms which use deep neural networks. Sadly, the reason behind the safety issues is also due to deep neural networks themselves. The pitfalls occur due to possible over-fitting and lack of awareness about the blind spots induced by the training distribution. Ideally, system designers would wish to cover as many scenarios during training as possible. However, achieving a meaningful coverage is impossible. This naturally leads to the following question: is it feasible to flag out-of-distribution (OOD) samples without causing too many false alarms? Such an OOD detector should be executable in a fashion that is computationally efficient. This is because OOD detectors often are executed as frequently as the sensors are sampled. Our aim in this article is to build an effective anomaly detector. To this end, we propose the idea of a memory bank to cache data samples which are representative enough to cover most of the in-distribution data. The similarity with respect to such samples can be a measure of familiarity of the test input. This is made possible by an appropriate choice of distance function tailored to the type of sensor we are interested in. Additionally, we adapt conformal anomaly detection framework to capture the distribution shifts with a guarantee of false alarm rate. We report the performance of our technique on two challenging scenarios: a self-driving car setting implemented inside the simulator CARLA with image inputs and autonomous racing car navigation setting with LiDAR inputs. From the experiments, it is clear that a deviation from the in-distribution setting can potentially lead to unsafe behavior. It should be noted that not all OOD inputs lead to precarious situations in practice, but staying in-distribution is akin to staying within a safety bubble and predictable behavior. An added benefit of our memory-based approach is that the OOD detector produces interpretable feedback for a human designer. This is of utmost importance since it recommends a potential fix for the situation as well. In other competing approaches, such feedback is difficult to obtain due to reliance on techniques which use variational autoencoders.

BabyBERTa, a language model trained on small-scale child-directed speech while none of the words are unmasked during training, has been shown to achieve a level of grammaticality comparable to that of RoBERTa-base, which is trained on 6,000 times more words and 15 times more parameters (Huebner et al., 2021).Relying on this promising result, we explore in this paper the performance of BabyBERTabased models in downstream tasks, focusing on Semantic Role Labeling (SRL) and two Extractive Question Answering tasks, with the aim of building more efficient systems that rely on less data and smaller models.We investigate the influence of these models both alone and as a starting point to larger pre-trained models, separately examining the contribution of the pre-training data, the vocabulary, and the masking policy on the downstream task performance.Our results show that BabyBERTa trained with unmasking-removal policy is a much stronger starting point for downstream tasks compared to the use of RoBERTa masking policy when 10M words are used for training and that this tendency persists, although to a lesser extent, when adding more training data. 1

Social interaction quality ratings derived from short natural conversations can differentiate children with and without autism at the group level. In this work, we explored conversations between children and an unfamiliar adult who rated their social interaction success on six dimensions. Using hand-crafted acoustic and lexical features, we built different classifiers to predict children's dimensional conversation quality. The best classifier achieved 61% accuracy, which outperformed human raters (49%). Follow-up analyses revealed that a subset of features determined communication quality scores. Additionally, we extracted acoustic features using a pretrained audio transformer and improved our prediction to 68%. This study suggests that automatically predicting conversation quality could be an inexpensive and objective way to monitor intervention progress in children with communication challenges, and could be used to identify intervention targets for improving conversational success.

Classifying high-dimensional hyperspectral image (HSI) with limited labeled samples is a difficult problem. One effective solution is to leverage knowledge from scenes with well-labeled image (the source domain) to aid training in the target domain. However, since the source and target domains have different category spaces, it is crucial to extract more discriminative features and address domain adaptation challenges. To tackle this issue, we propose a cross-scene classification method for HSIs via generative adversarial networks (GANs) in latent space (GLS). Our method employs autoencoders (AEs) to map the input data to a latent space, where the most effective feature representation is extracted and preserved by deep residual 3D convolutional neural networks (CNN). The unlabeled samples in the target domain are also utilized in the AE which ensure all the samples are considered. We leverage conditional adversarial domain adaptation to overcome the domain shift, and introduce maximum mean discrepancy loss to minimize distribution differences between the two domains, facilitating better domain distribution alignment. We tested our approach on three public datasets and demonstrated that it outperforms existing few-shot learning methods. Our results highlight the effectiveness of our classification method via GANs in latent space for HSIs, and show that it has potential for practical applications.

Abstract Symbolic model checking is an important tool for finding bugs (or proving the absence of bugs) in modern system designs. Because of this, improving the ease of use, scalability, and performance of model checking tools and algorithms continues to be an important research direction. In service of this goal, we present , an open-source SMT-based model checker. is designed to be both a research platform for developing and improving model checking algorithms, as well as a performance-competitive tool that can be used for academic and industry verification applications. In addition to performance, prioritizes transparency (developed as an open-source project on GitHub), flexibility ( can be adapted to a variety of tasks by exploiting its general SMT-based interface), and extensibility (it is easy to add new algorithms and new back-end solvers). In this paper, we describe the design of the tool with a focus on the flexible and extensible architecture, cover its current capabilities, and demonstrate that is competitive with state-of-the-art tools.

As designs grow in size and complexity, design verification becomes one of the most difficult and costly tasks facing design teams. Formal verification techniques offer great promise because of their ability to exhaustively explore design behaviors. However, formal techniques also have a reputation for being labor-intensive and limited to small blocks. Is there any hope for successful application of formal techniques at design scale? We answer this question affirmatively by digging deeper to understand what the real technological issues and opportunities are. First, we look at satisfiability solvers, the engines underlying formal techniques such as model checking. Given the recent innovations in satisfiability solving, we argue that there are many reasons to be optimistic that formal techniques will scale to designs of practical interest. We use our CoSA model checker as a demonstration platform to illustrate how advances in solvers can improve scalability. However, even if solvers become blazingly fast, applying them well is still labor-intensive. This is because formal tools are only as useful as the properties they are given to prove, which traditionally have required great effort to develop. Symbolic quick error detection (SQED) addresses this issue by using a single, universal property that checks designs automatically. We demonstrate how SQED can automatically find logic and security bugs in a variety of designs and report on bugs found and efficiency gains realized in academic and industry designs. We also present a generator for an improved SQED module that further reduces the amount of manual effort that has to be spent by the designer.