Abstract

Automatic detection and classification of stuttering events remain a challenging problem in speech processing, particularly due to the variability of disfluency types and the scarcity of large annotated datasets. This paper presents a signal processing-based approach for stuttering classification that avoids the data-intensive requirements of machine learning models. Our method focuses on syllable-level analysis using key acoustic parameters such as energy patterns, pitch contours, spectral stability, and temporal features, implemented through rule-based classification with predefined thresholds. The methodology involves automatic syllable segmentation followed by rule-based classification using thresholds on acoustic features. The dataset comprises annotated read and spontaneous speech from 106 Kannada-speaking adults, including 26 new participants evaluated during clinical trials. Three certified Speech-Language Pathologists (SLPs) conducted perceptual evaluations, and system outputs were validated using intraclass correlation (ICC), yielding strong agreement (textgreater 0.82) across stuttering types. The system achieved classification accuracies of 89% for blocks, 83% for repetitions, and 81% for prolongations. However, the system has several limitations: it relies on static threshold values, exhibits sensitivity to noise, and lacks automated severity grading. Additionally, generalizability across languages and speaking conditions requires further exploration. Machine learning models such as SVM, TDNN, and LSTM were trained and achieved a performance of around 30 to 60% accuracy. Our approach offers high interpretability, low computational cost, and real-time feasibility compared to machine learning methods. Future work will focus on adaptive thresholding, noise-robust processing, and hybrid approaches integrating signal processing with lightweight learning-based refinement.

Abstract

Toxicity identification in online multimodal environments remains a challenging task due to the complexity of contextual connections across modalities (e.g., textual and visual). In this paper, we propose a novel framework that integrates Knowledge Distillation (KD) from Large Visual Language Models (LVLMs) and knowledge infusion to enhance the performance of toxicity detection in hateful memes. Our approach extracts sub-knowledge graphs from ConceptNet, a large-scale commonsense Knowledge Graph (KG) to be infused within a compact VLM framework. The relational context between toxic phrases in captions and memes, as well as visual concepts in memes enhance the model's reasoning capabilities. Experimental results from our study on two hate speech benchmark datasets demonstrate superior performance over the state-of-the-art baselines across AU-ROC, F1, and Recall with improvements of 1.1%, 7%, and 35%, respectively. Given the contextual complexity of the toxicity detection task, our approach showcases the significance of learning from both explicit (i.e. KG) as well as implicit (i.e. LVLMs) contextual cues incorporated through a hybrid neurosymbolic approach. This is crucial for real-world applications where accurate and scalable recognition of toxic content is critical for creating safer online environments.

Abstract

Federated Learning (FL) enables multiple users to collaboratively train a global model in a distributed manner without revealing their personal data. However, FL remains vulnerable to model poisoning attacks, where malicious actors inject crafted updates to compromise the global model’s accuracy. We propose a novel defense mechanism, Kernel-based Trust Segmentation (KeTS), to counter model poisoning attacks. Unlike existing approaches, KeTS analyzes the evolution of each client’s updates and effectively segments malicious clients using Kernel Density Estimation (KDE), even in the presence of benign outliers. We thoroughly evaluate KeTS’s performance against the six most effective model poisoning attacks (i.e., Trim-Attack, Krum-Attack, Min-Max attack, Min-Sum attack, and their variants) on four different datasets (i.e., MNIST, Fashion-MNIST, CIFAR-10, and KDD-CUP-1999) and compare its performance with three classical robust schemes (i.e., Krum, TrimMean, and Median) and a state-of-the-art defense (i.e., FLTrust). Our results show that KeTS outperforms the existing defenses in every attack setting; beating the best-performing defense by an overall average of >24% (on MNIST), >14% (on Fashion-MNIST), >9% (on CIFAR-10), > 11% (on KDD-CUP-1999). A series of further experiments (varying poisoning approaches, attacker population, etc.) reveal the consistent and superior performance of KeTS under diverse conditions. KeTS is a practical solution as it satisfies all three defense objectives (i.e., fidelity, robustness, and efficiency) without imposing additional overhead on the clients. Finally, we also discuss a simple, yet effective extension to KeTS to handle consistent-untargeted (e.g., sign-flipping) attacks as well as targeted attacks (e.g., label-flipping.

Abstract

This paper presents the submission of IIITHBUT to the IWSLT 2025 shared task on speech translation for the low-resource Bhojpuri-Hindi language pair. We explored the impact of hyperparameter optimisation and data augmentation techniques on the performance of the SeamlessM4T model fine-tuned for this specific task. We systematically investigated a range of hyperparameters including learning rate schedules, number of update steps, warm-up steps, label smoothing, and batch sizes; and report their effect on translation quality. To address data scarcity, we applied speed perturbation and SpecAugment and studied their effect on translation quality. We also examined the use of cross-lingual signal through joint training with Marathi and Bhojpuri speech data. Our experiments reveal that careful selection of hyperparameters and the application of simple yet effective augmentation techniques significantly improve performance in low-resource settings. We also analysed the translation hypotheses to understand various kinds of errors that impacted the translation quality in terms of BLEU

Abstract

Exploring promiscuous catalytic activity of enzymes to stereoselective transformations uncovers ecofriendly asymmetric catalysts and provides scope to execute new-to-nature chemistry. Biocatalytic promiscuous diastereoselective Henry reaction (DHR) is restricted to a few hydroxynitrile lyases (HNLs), often engineered, with narrow synthetic scope and anti-selectivity. Here, we report a single native enzyme exhibiting complementary diastereoselectivity in an asymmetric Henry reaction. Baliospermum montanum HNL (BmHNL)-catalyzed DHR enabled the production of thirty-two anti-(1S,2R)-β-nitroalcohols with up to >99% conversion, >99% ee, and >99% de using longer nitroalkanes. It afforded syn diastereomers, i.e., (1S,2S)-β-nitroalcohols as thermodynamically controlled products by manipulating the reaction conditions. Reuse of BmHNL in DHR for 20 cycles has empowered the synthesis of (1S,2R)-2-nitro-1-phenylpropan-1-ol (NPP), where it retained 82% of its initial activity, maintained >99% ee on each cycle, and provided total turnover number >3.7 × 104, which is 140-fold higher compared to the only existing biocatalytic DHR. Our study demonstrates a gram-scale synthesis of (1S,2R)-NPP and a preparative-scale synthesis of (1S,2S)-NPP, a precursor to d-norpseudoephedrine used for the treatment of obesity. The origin of the inverse diastereoselectivity was probed using isotope labeling studies. Molecular modeling of the reaction using density functional theory methods reveals the underlying mechanism of the diastereoselectivity and the competing kinetic vs thermodynamic nature of the reaction. Despite a promiscuous catalytic activity, the outstanding catalytic efficiency, broad synthetic scope, and complementary diastereoselectivity of the native enzyme of BmHNL on DHR are remarkable, which illustrates it as a specialized biocatalyst for greener technology and stereocontrolled production of diverse β-nitroalcohols.

Abstract

Evaluating tables qualitatively and quantitatively poses a significant challenge, as standard metrics often overlook subtle structural and content-level discrepancies. To address this, we propose a rubric-based evaluation framework that integrates multi-level structural descriptors with fine-grained contextual signals, enabling more precise and consistent table comparison. Building on this, we introduce TabXEval, an eXhaustive and eXplainable two-phase evaluation framework. TabXEval first aligns reference and predicted tables structurally via TabAlign, then performs semantic and syntactic comparison using TabCompare, offering interpretable and granular feedback. We evaluate TabXEval on TabXBench, a diverse, multi-domain benchmark featuring realistic table perturbations and human annotations. A sensitivity-specificity analysis further demonstrates the robustness and explainability of TabXEval across varied table tasks. Code and data are available at https://coral-lab-asu.github.io/tabxeval/.

Abstract

Cataract surgery is the most common surgical procedure globally, with a disproportionately higher burden in developing countries. While automated surgical video analysis has been explored in general surgery, its application to ophthalmic procedures remains limited. Existing research primarily focuses on Phaco cataract surgery, an expensive technique not accessible in regions where cataract treatment is most needed. In contrast, Manual Small-Incision Cataract Surgery (MSICS) is the preferred low-cost, faster alternative in high-volume settings and for complex cases. However, no dataset exists for MSICS. To address this gap, we introduce Cataract-MSICS, the first comprehensive dataset containing 53 surgical videos annotated for 18 surgical phases and 3,527 frames with 13 surgical tools at the pixel level. We also present ToolSeg, a novel framework that enhances tool segmentation with a phase-conditional decoder and a semi-supervised setup leveraging pseudo-labels from foundation models. Our approach significantly improves segmentation performance, achieving a 38.10% increase in mean Dice scores, with notable gains for less prevalent and smaller tools.

Abstract

Maritime communication, critical for global oceanic trade, faces challenges and opportunities with advancements in Information and Communication Technology (ICT). Traditional methods are susceptible to interception due to open channels, limited authentication, jamming, and other security risks. Securing vessel movements and locations in Internet of Vessels (IoV) is essential to prevent unauthorized data interception and tampering during transmission. We propose a ship authentication method using location-based secure keys to ensure the confidentiality of a vessel’s whereabouts. The proposed scheme’s robustness is validated through formal and informal security analyses, and formal verification using the Scyther automated verification tool, demonstrating its effectiveness against potential attacks in maritime networks. Comparative studies indicate that utilizing location-based keys maintains anonymity and untraceability without imposing significant computational or communication burdens. Experimental findings from comprehensive big data analytics and simulations using NS3 validate the scheme’s feasibility and performance.

Abstract

Community detection is the problem of identifying natural divisions in networks. Efficient parallel algorithms for this purpose are crucial in various applications, particularly as datasets grow to substantial scales. This paper presents an optimized parallel implementation of the Label Propagation Algorithm (LPA), a high speed community detection method, for shared memory multicore systems. On a server equipped with dual 16-core Intel Xeon Gold 6226R processors, our LPA, which we term as GVE-LPA, outperforms FLPA, igraph LPA, and NetworKit LPA by 139x, 97000x, and 40x respectively - achieving a processing rate of 1.4B edges/s on a 3.8B edge graph. In addition, GVE-LPA scales at a rate of 1.7x every doubling of threads.

Abstract

Natural ventilation (NV) is vital for optimizing energy efficiency, particularly when outdoor conditions are favourable. In highly polluted regions like National Capital Regions of Delhi (Delhi-NCR), occupants often avoid NV during pollution episodes, relying instead on mechanical systems. This behaviour stems from the assumption that outdoor air is consistently unsuitable for ventilation, even when pollution levels are within acceptable limits. As a result, energy consumption increases, even during periods when outdoor air could support NV. This study integrates PM2.5 and PM10 data to develop a pollution-aware NV strategy for residential buildings. Behavioural insights gathered from detailed ethnographic interviews in Delhi-NCR highlight occupant tendencies to keep windows closed during pollution episodes, often based on perceived air quality rather than actual data. While previous studies have examined the effects of pollution on NV, few have directly linked measured air quality data with real occupant behaviour in residential settings. The proposed framework addresses these behavioural patterns. If outdoor PM2.5 and PM10 levels exceed standard thresholds, occupants are advised to close windows to maintain indoor air quality. Alternatively, when these levels fall below predefined thresholds, occupants are encouraged to open windows, highlighting the potential for energy savings. For the above-mentioned scenarios building energy simulation was employed to simulate NV strategies in a mixed mode building. This pollution-aware approach bridges the gap between occupant behaviour and actual environmental conditions, promoting energy-efficient ventilation practices in regions affected by air pollution. Our results demonstrate that aligning behavioural NV strategies with actual data addresses air quality concerns, contributing to energy-efficient and occupant-centric ventilation solutions.