Abstract

Curriculum learning has been used to improve the quality of text generation systems by ordering the training samples according to a particular schedule in various tasks. In the context of data-to-text generation (DTG), previous studies used various difficulty criteria to order the training samples for monolingual DTG. These criteria, however, do not generalize to the crosslingual variant of the problem and do not account for noisy data. We explore multiple criteria that can be used for improving the performance of cross-lingual DTG systems with noisy data using two curriculum schedules. Using the alignment score criterion for ordering samples and an annealing schedule to train the model, we show increase in BLEU score by up to 4 points, and improvements in faithfulness and coverage of generations by 5-15% on average across 11 Indian languages and English in 2 separate datasets. We make code and data publicly available

Abstract

We introduce a family of information leakage measures called maximal (α,β) -leakage ( Mα beL), parameterized by real numbers α and β greater than or equal to 1. The measure is formalized via an operational definition involving an adversary guessing an unknown (randomized) function of the data given the released data. We obtain a simplified computable expression for the measure and show that it satisfies several basic properties such as monotonicity in β for a fixed α , non-negativity, data processing inequalities, and additivity over independent releases. We highlight the relevance of this family by showing that it bridges several known leakage measures, including maximal α -leakage (β=1) , maximal leakage (α=∞,β=1) , local differential privacy (LDP) (α=∞,β=∞) , and local Rényi differential privacy (LRDP) (α=β) , thereby giving an operational interpretation to local Rényi differential privacy. We also study a conditional version of Mα beL on leveraging which we recover differential privacy and Rényi differential privacy. A new variant of LRDP, which we call maximal Rényi leakage, appears as a special case of Mα beL for α=∞ that smoothly tunes between maximal leakage ( β=1 ) and LDP ( β=∞ ). Finally, we show that a vector form of the maximal Rényi leakage relaxes differential privacy under Gaussian and Laplacian mechanisms.

Abstract

Generative adversarial networks (GANs), modeled as a zero-sum game between a generator (G) and a discriminator (D), allow generating synthetic data with formal guarantees. Noting that D is a classifier, we begin by reformulating the GAN value function using class probability estimation (CPE) losses. We prove a two-way correspondence between CPE loss GANs and f-GANs which minimize f-divergences. We also show that all symmetric f-divergences are equivalent in convergence. In the finite sample and model capacity setting, we define and obtain bounds on estimation and generalization errors. We specialize these results to α -GANs, defined using α -loss, a tunable CPE loss family parametrized by α∈(0,∞ ]. We next introduce a class of dual-objective GANs to address training instabilities of GANs by modeling each player’s objective using α -loss to obtain (αD,αG) -GANs. We show that the resulting non-zero sum game simplifies to minimizing an f-divergence under appropriate conditions on (αD,αG) . Generalizing this dual-objective formulation using CPE losses, we define and obtain upper bounds on an appropriately defined estimation error. Finally, we highlight the value of tuning (αD,αG) in alleviating training instabilities for the synthetic 2D Gaussian mixture ring as well as the large publicly available Celeb-A and LSUN Classroom image datasets.

Abstract

Distracted driving is a leading cause of road accidents globally. Identifying distracted driving involves reliably detecting and classifying various forms of driver distraction (e.g., texting, eating, or using in-car devices) from in-vehicle camera feeds to enhance road safety. This task is challenging due to the need for robust models that can generalize to a diverse set of driver behaviors without requiring extensive annotated datasets. In this paper, we propose KiD3, a novel method for distracted driver detection (DDD) that infuses auxiliary knowledge about semantic relations between entities in a scene and the structural configuration of the driver’s pose. Specifically, we construct a unified framework that integrates scene graphs and driver pose information with visual cues from video frames to create a holistic representation of the driver’s actions. Our results indicate that KiD3 achieves a 13.64% accuracy improvement over the vision-only baseline by incorporating auxiliary knowledge with visual information.

Abstract

Large Language Models (LLMs) have emerged as powerful tools to per- form various tasks in the legal domain, ranging from generating summaries to pre- dicting judgments. Despite their immense potential, these models have been proven to learn and exhibit societal biases and make unfair predictions. Hence, it is es- sential to evaluate these models prior to deployment. In this study, we explore the ability of LLMs to perform Binary Statutory Reasoning in the Indian legal land- scape across various societal disparities. We present a novel metric, β -weighted Le- gal Safety Score (LSSβ ), to evaluate the legal usability of the LLMs. Additionally, we propose a finetuning pipeline, utilising specialised legal datasets, as a potential method to reduce bias. Our proposed pipeline effectively reduces bias in the model, as indicated by improved LSSβ . This highlights the potential of our approach to enhance fairness in LLMs, making them more reliable for legal tasks in socially diverse contexts.

Abstract

This paper uses 3D stochastic geometry to analyse the coverage and meta-distribution (MD) performance of a low Earth orbit (LEO) satellite-based Internet-of-Things (IoT) network. The satellites are considered to be distributed at a fixed altitude around Earth following a Binomial point process. An IoT device broadcasts its sensed information to all the satellites with a line-of-sight within the visibility region and they act as selective decode-and-forward gateways. The ground server coherently combines the information received from multiple satellites. The analytically derived performance measures are verified through Monte Carlo simulations. The results demonstrate the impact of mega-LEO constellations on coverage and reliability, guiding 6G architecture design to improve connectivity and data offloading in smart cities and dense urban environments.

Abstract

A visual question answering (VQA) system for electrical circuit images could be useful as a quiz generator, design and verification assistant or an electrical diagnosis tool. Although there exists a vast literature on VQA, to the best of our knowledge, there is no existing work on VQA for electrical circuit images. To this end, we curate a new dataset, CircuitVQA, of 115K+ questions on 5725 electrical images with 70 circuit symbols. The dataset contains schematic as well as hand-drawn images. The questions span various categories like counting, value, junction and position based questions. To be effective, models must demonstrate skills like object detection, text recognition, spatial understanding, question intent understanding and answer generation. We experiment with multiple foundational visio-linguistic models for this task and find that a finetuned BLIP model with component descriptions as additional input provides best results. We make the code and dataset publicly available

Abstract

In literature, various time-frequency representation methods were investigated for automatic detection of voice disorders. Stockwell-Transform (S-Transform) provides good time-frequency localization; hence, it may efficiently capture the voice disorder related information from speech signal. With this motivation, we investigated different variants of S-Transform for the classification of voice disorders. This study proposed the S-Transform based cepstral coefficients for voice disorder detection and assessment. The performance of the proposed feature was compared with baseline features on SVD and HUPA databases. Compared to baseline features, proposed features performed best in terms of classification accuracy of 80.2% and 79.8% on HUPA and SVD databases, respectively for voice disorder detection task. Also, the proposed features performed better in case of assessment task. Further, the experimental results reveal that combining cepstral coefficients derived from S-Transform with baseline features improved the performance of proposed systems by 8% and 4% for detection and assessment tasks, respectively which highlights complementary nature of the explored features. We also analysed the effectiveness of S-Transform based spectral representation in capturing the acoustic characteristics for various voice qualities like breathiness, harshness, creakiness, and falsetto phonations. This representation was also compared with other time-frequency based methods such as STFT, ZTW and SFF. It was observed that S-Transform effectively captures the acoustic variations associated with different voice qualities compared to other baseline methods, which may be due to better spectro-temporal resolution offered by the S-Transform.

Abstract

3D Gaussian Splatting (3DGS) has transformed novel-view synthesis with its fast, interpretable, and high-fidelity rendering. However, its resource requirements limit its usability. Especially on constrained devices, training performance degrades quickly and often cannot complete due to excessive memory consumption of the model. The method converges with an indefinite number of Gaussians—many of them redundant—making rendering unnecessarily slow and preventing its usage in downstream tasks that expect fixed-size inputs. To address these issues, we tackle the challenges of training and rendering 3DGS models on a budget. We use a guided, purely constructive densification process that steers densification toward Gaussians that raise the reconstruction quality. Model size continuously increases in a controlled manner towards an exact budget, using score-based densification of Gaussians with training-time priors that measure their contribution. We further address training speed obstacles: following a careful analysis of 3DGS’ original pipeline, we derive faster, numerically equivalent solutions for gradient computation and attribute updates, including an alternative parallelization for efficient backpropagation. We also propose quality-preserving approximations where suitable to reduce training time even further. Taken together, these enhancements yield a robust, scalable solution with reduced training times, lower compute and memory requirements, and high quality. Our evaluation shows that in a budgeted setting, we obtain competitive quality metrics with 3DGS while achieving a 4–5 × reduction in both model size and training time. With more generous budgets, our measured quality surpasses theirs. These advances open the door for novel-view synthesis in constrained environments, e.g., mobile devices.

Abstract

The IEEE International Conference on Software Architecture (ICSA) is the premier gathering of practitioners and researchers interested in software architecture, component-based software engineering, and quality aspects of complex software systems. The 21st IEEE International Conference on Software Architecture (ICSA 2024) continued the tradition of a working conference, where attendees met and where software architects were able to explain the challenges they face and try to influence the future of the field. Interactive working sessions were the place where researchers met practitioners to identify opportunities to shape the future of our field.