In this paper, we explore the properties of loss curvature with respect to input data in deep neural networks. Curvature of loss with respect to input (termed input loss curvature) is the trace of the Hessian of the loss with respect to the input. We investigate how input loss curvature varies between train and test sets, and its implications for train-test distinguishability. We develop a theoretical framework that derives an upper bound on the train-test distinguishability based on privacy and the size of the training set. This novel insight fuels the development of a new black box membership inference attack utilizing input loss curvature. We validate our theoretical findings through experiments in computer vision classification tasks, demonstrating that input loss curvature surpasses existing methods in membership inference effectiveness. Our analysis highlights how the performance of membership inference attack (MIA) methods varies with the size of the training set, showing that curvature-based MIA outperforms other methods on sufficiently large datasets. This condition is often met by real datasets, as demonstrated by our results on CIFAR10, CIFAR100, and ImageNet. These findings not only advance our understanding of deep neural network behavior but also improve the ability to test privacy-preserving techniques in machine learning.

Context: Machine learning (ML) and deep learning (DL) analyze raw data to extract valuable insights in specific phases. The rise of continuous practices in software projects emphasizes automating Continuous Integration (CI) with these learning-based methods, while the growing adoption of such approaches underscores the need for systematizing knowledge. Objective: Our objective is to comprehensively review and analyze existing literature concerning learning-based methods within the CI domain. We endeavour to identify and analyse various techniques documented in the literature, emphasizing the fundamental attributes of training phases within learning-based solutions in the context of CI. Method: We conducted a Systematic Literature Review (SLR) involving 52 primary studies. Through statistical and thematic analyses, we explored the correlations between CI tasks and the training phases of learning-based methodologies across the selected studies, encompassing a spectrum from data engineering techniques to evaluation metrics. Results: This paper presents an analysis of the automation of CI tasks utilizing learning-based methods. We identify and analyze nine types of data sources, four steps in data preparation, four feature types, nine subsets of data features, five approaches for hyperparameter selection and tuning, and fifteen evaluation metrics. Furthermore, we discuss the latest techniques employed, existing gaps in CI task automation, and the characteristics of the utilized learning-based techniques. Conclusion: This study provides a comprehensive overview of learning-based methods in CI, offering valuable insights for researchers and practitioners developing CI task automation. It also highlights the need for further research to advance these methods in CI.

Deepfake technology is rapidly advancing, posing significant challenges to the detection of manipulated media content. Parallel to that, some adversarial attack techniques have been developed to fool the deepfake detectors and make deepfakes even more difficult to be detected. This paper explores the application of super resolution techniques as a possible adversarial attack in deepfake detection. Through our experiments, we demonstrate that minimal changes made by these methods in the visual appearance of images can have a profound impact on the performance of deepfake detection systems. We propose a novel attack using super resolution as a quick, black-box and effective method to camouflage fake images and/or generate false alarms on pristine images. Our results indicate that the usage of super resolution can significantly impair the accuracy of deepfake detectors, thereby highlighting the vulnerability of such systems to adversarial attacks.

Parameter-efficient fine-tuning (PEFT) for personalizing automatic speech recognition (ASR) has recently shown promise for adapting general population models to atypical speech. However, these approaches assume a priori knowledge of the atypical speech disorder being adapted for -- the diagnosis of which requires expert knowledge that is not always available. Even given this knowledge, data scarcity and high inter/intra-speaker variability further limit the effectiveness of traditional fine-tuning. To circumvent these challenges, we first identify the minimal set of model parameters required for ASR adaptation. Our analysis of each individual parameter's effect on adaptation performance allows us to reduce Word Error Rate (WER) by half while adapting 0.03% of all weights. Alleviating the need for cohort-specific models, we next propose the novel use of a meta-learned hypernetwork to generate highly individualized, utterance-level adaptations on-the-fly for a diverse set of atypical speech characteristics. Evaluating adaptation at the global, cohort and individual-level, we show that hypernetworks generalize better to out-of-distribution speakers, while maintaining an overall relative WER reduction of 75.2% using 0.1% of the full parameter budget.

Differential privacy (DP) is a widely used approach for mitigating privacy risks when training machine learning models on sensitive data. DP mechanisms add noise during training to limit the risk of information leakage. The scale of the added noise is critical, as it determines the trade-off between privacy and utility. The standard practice is to select the noise scale in terms of a privacy budget parameter $\epsilon$. This parameter is in turn interpreted in terms of operational attack risk, such as accuracy, or sensitivity and specificity of inference attacks against the privacy of the data. We demonstrate that this two-step procedure of first calibrating the noise scale to a privacy budget $\epsilon$, and then translating $\epsilon$ to attack risk leads to overly conservative risk assessments and unnecessarily low utility. We propose methods to directly calibrate the noise scale to a desired attack risk level, bypassing the intermediate step of choosing $\epsilon$. For a target attack risk, our approach significantly decreases noise scale, leading to increased utility at the same level of privacy. We empirically demonstrate that calibrating noise to attack sensitivity/specificity, rather than $\epsilon$, when training privacy-preserving ML models substantially improves model accuracy for the same risk level. Our work provides a principled and practical way to improve the utility of privacy-preserving ML without compromising on privacy.

Quantifying the uncertainty of predictions made by large language models (LLMs) in binary text classification tasks remains a challenge. Calibration, in the context of LLMs, refers to the alignment between the model's predicted probabilities and the actual correctness of its predictions. A well-calibrated model should produce probabilities that accurately reflect the likelihood of its predictions being correct. We propose a novel approach that utilizes the inductive Venn--Abers predictor (IVAP) to calibrate the probabilities associated with the output tokens corresponding to the binary labels. Our experiments on the BoolQ dataset using the Llama 2 model demonstrate that IVAP consistently outperforms the commonly used temperature scaling method for various label token choices, achieving well-calibrated probabilities while maintaining high predictive quality. Our findings contribute to the understanding of calibration techniques for LLMs and provide a practical solution for obtaining reliable uncertainty estimates in binary question answering tasks, enhancing the interpretability and trustworthiness of LLM predictions.

Offering a promising solution to the scalability challenges associated with human evaluation, the LLM-as-a-judge paradigm is rapidly gaining traction as an approach to evaluating large language models (LLMs). However, there are still many open questions about the strengths and weaknesses of this paradigm, and what potential biases it may hold. In this paper, we present a comprehensive study of the performance of various LLMs acting as judges. We leverage TriviaQA as a benchmark for assessing objective knowledge reasoning of LLMs and evaluate them alongside human annotations which we found to have a high inter-annotator agreement. Our study includes 9 judge models and 9 exam taker models -- both base and instruction-tuned. We assess the judge model's alignment across different model sizes, families, and judge prompts. Among other results, our research rediscovers the importance of using Cohen's kappa as a metric of alignment as opposed to simple percent agreement, showing that judges with high percent agreement can still assign vastly different scores. We find that both Llama-3 70B and GPT-4 Turbo have an excellent alignment with humans, but in terms of ranking exam taker models, they are outperformed by both JudgeLM-7B and the lexical judge Contains, which have up to 34 points lower human alignment. Through error analysis and various other studies, including the effects of instruction length and leniency bias, we hope to provide valuable lessons for using LLMs as judges in the future.

X-ray image-based disease diagnosis lies in ensuring the precision of identifying afflictions within the sample, a task fraught with challenges stemming from the occurrence of false positives and false negatives. False positives introduce the risk of erroneously identifying non-existent conditions, leading to misdiagnosis and a decline in patient care quality. Conversely, false negatives pose the threat of overlooking genuine abnormalities, potentially causing delays in treatment and interventions, thereby resulting in adverse patient outcomes. The urgency to overcome these challenges compels ongoing efforts to elevate the precision and reliability of X-ray image analysis algorithms within the computational framework. This study introduces modified pre-trained ResNet models tailored for multi-class disease diagnosis of X-ray images, incorporating advanced optimization strategies to reduce the execution runtime of training and inference tasks. The primary objective is to achieve tangible performance improvements through accelerated implementations of PyTorch, CUDA, Mixed- Precision Training, and Learning Rate Scheduler. While outcomes demonstrate substantial improvements in execution runtimes between normal training and CUDA-accelerated training, negligible differences emerge between various training optimization modalities. This research marks a significant advancement in optimizing computational approaches to reduce training execution time for larger models. Additionally, we explore the potential of effective parallel data processing using MPI4Py for the distribution of gradient descent optimization across multiple nodes and leverage multiprocessing to expedite data preprocessing for larger datasets.

The single-stage point-based 3D object detectors have attracted widespread research interest due to their advantages of lightweight and fast inference speed. However, they still face challenges such as inadequate learning of low-quality objects (ILQ) and misalignment between localization accuracy and classification confidence (MLC). In this paper, we propose SGCCNet to alleviate these two issues. For ILQ, SGCCNet adopts a Saliency-Guided Data Augmentation (SGDA) strategy to enhance the robustness of the model on low-quality objects by reducing its reliance on salient features. Specifically, We construct a classification task and then approximate the saliency scores of points by moving points towards the point cloud centroid in a differentiable process. During the training process, SGCCNet will be forced to learn from low saliency features through dropping points. Meanwhile, to avoid internal covariate shift and contextual features forgetting caused by dropping points, we add a geometric normalization module and skip connection block in each stage. For MLC, we design a Confidence Correction Mechanism (CCM) specifically for point-based multi-class detectors. This mechanism corrects the confidence of the current proposal by utilizing the predictions of other key points within the local region in the post-processing stage. Extensive experiments on the KITTI dataset demonstrate the generality and effectiveness of our SGCCNet. On the KITTI \textit{test} set, SGCCNet achieves $80.82\%$ for the metric of $AP_{3D}$ on the \textit{Moderate} level, outperforming all other point-based detectors, surpassing IA-SSD and Fast Point R-CNN by $2.35\%$ and $3.42\%$, respectively. Additionally, SGCCNet demonstrates excellent portability for other point-based detectors

Microvascular anatomy is known to be involved in various neurological disorders. However, understanding these disorders is hindered by the lack of imaging modalities capable of capturing the comprehensive three-dimensional vascular network structure at microscopic resolution. With a lateral resolution of $<=$20 {\textmu}m and ability to reconstruct large tissue blocks up to tens of cubic centimeters, serial-section optical coherence tomography (sOCT) is well suited for this task. This method uses intrinsic optical properties to visualize the vessels and therefore does not possess a specific contrast, which complicates the extraction of accurate vascular models. The performance of traditional vessel segmentation methods is heavily degraded in the presence of substantial noise and imaging artifacts and is sensitive to domain shifts, while convolutional neural networks (CNNs) require extensive labeled data and are also sensitive the precise intensity characteristics of the data that they are trained on. Building on the emerging field of synthesis-based training, this study demonstrates a synthesis engine for neurovascular segmentation in sOCT images. Characterized by minimal priors and high variance sampling, our highly generalizable method tested on five distinct sOCT acquisitions eliminates the need for manual annotations while attaining human-level precision. Our approach comprises two phases: label synthesis and label-to-image transformation. We demonstrate the efficacy of the former by comparing it to several more realistic sets of training labels, and the latter by an ablation study of synthetic noise and artifact models.