Sub-word segmentation is an essential pre-processing step for Neural Machine Translation (NMT). Existing work has shown that neural sub-word segmenters are better than Byte-Pair Encoding (BPE), however, they are inefficient as they require parallel corpora, days to train and hours to decode. This paper introduces SelfSeg, a self-supervised neural sub-word segmentation method that is much faster to train/decode and requires only monolingual dictionaries instead of parallel corpora. SelfSeg takes as input a word in the form of a partially masked character sequence, optimizes the word generation probability and generates the segmentation with the maximum posterior probability, which is calculated using a dynamic programming algorithm. The training time of SelfSeg depends on word frequencies, and we explore several word frequency normalization strategies to accelerate the training phase. Additionally, we propose a regularization mechanism that allows the segmenter to generate various segmentations for one word. To show the effectiveness of our approach, we conduct MT experiments in low-, middle- and high-resource scenarios, where we compare the performance of using different segmentation methods. The experimental results demonstrate that on the low-resource ALT dataset, our method achieves more than 1.2 BLEU score improvement compared with BPE and SentencePiece, and a 1.1 score improvement over Dynamic Programming Encoding (DPE) and Vocabulary Learning via Optimal Transport (VOLT) on average. The regularization method achieves approximately a 4.3 BLEU score improvement over BPE and a 1.2 BLEU score improvement over BPE-dropout, the regularized version of BPE. We also observed significant improvements on IWSLT15 Vi->En, WMT16 Ro->En and WMT15 Fi->En datasets, and competitive results on the WMT14 De->En and WMT14 Fr->En datasets.

Self-supervised learning on graphs has made large strides in achieving great performance in various downstream tasks. However, many state-of-the-art methods suffer from a number of impediments, which prevent them from realizing their full potential. For instance, contrastive methods typically require negative sampling, which is often computationally costly. While non-contrastive methods avoid this expensive step, most existing methods either rely on overly complex architectures or dataset-specific augmentations. In this paper, we ask: Can we borrow from classical unsupervised machine learning literature in order to overcome those obstacles? Guided by our key insight that the goal of distance-based clustering closely resembles that of contrastive learning: both attempt to pull representations of similar items together and dissimilar items apart. As a result, we propose CARL-G - a novel clustering-based framework for graph representation learning that uses a loss inspired by Cluster Validation Indices (CVIs), i.e., internal measures of cluster quality (no ground truth required). CARL-G is adaptable to different clustering methods and CVIs, and we show that with the right choice of clustering method and CVI, CARL-G outperforms node classification baselines on 4/5 datasets with up to a 79x training speedup compared to the best-performing baseline. CARL-G also performs at par or better than baselines in node clustering and similarity search tasks, training up to 1,500x faster than the best-performing baseline. Finally, we also provide theoretical foundations for the use of CVI-inspired losses in graph representation learning.

Vulnerability detection is a critical problem in software security and attracts growing attention both from academia and industry. Traditionally, software security is safeguarded by designated rule-based detectors that heavily rely on empirical expertise, requiring tremendous effort from software experts to generate rule repositories for large code corpus. Recent advances in deep learning, especially Graph Neural Networks (GNN), have uncovered the feasibility of automatic detection of a wide range of software vulnerabilities. However, prior learning-based works only break programs down into a sequence of word tokens for extracting contextual features of codes, or apply GNN largely on homogeneous graph representation (e.g., AST) without discerning complex types of underlying program entities (e.g., methods, variables). In this work, we are one of the first to explore heterogeneous graph representation in the form of Code Property Graph and adapt a well-known heterogeneous graph network with a dual-supervisor structure for the corresponding graph learning task. Using the prototype built, we have conducted extensive experiments on both synthetic datasets and real-world projects. Compared with the state-of-the-art baselines, the results demonstrate promising effectiveness in this research direction in terms of vulnerability detection performance (average F1 improvements over 10\% in real-world projects) and transferability from C/C++ to other programming languages (average F1 improvements over 11%).

The demand for artificial intelligence (AI) in healthcare is rapidly increasing. However, significant challenges arise from data scarcity and privacy concerns, particularly in medical imaging. While existing generative models have achieved success in image synthesis and image-to-image translation tasks, there remains a gap in the generation of 3D semantic medical images. To address this gap, we introduce Med-DDPM, a diffusion model specifically designed for semantic 3D medical image synthesis, effectively tackling data scarcity and privacy issues. The novelty of Med-DDPM lies in its incorporation of semantic conditioning, enabling precise control during the image generation process. Our model outperforms Generative Adversarial Networks (GANs) in terms of stability and performance, generating diverse and anatomically coherent images with high visual fidelity. Comparative analysis against state-of-the-art augmentation techniques demonstrates that Med-DDPM produces comparable results, highlighting its potential as a data augmentation tool for enhancing model accuracy. In conclusion, Med-DDPM pioneers 3D semantic medical image synthesis by delivering high-quality and anatomically coherent images. Furthermore, the integration of semantic conditioning with Med-DDPM holds promise for image anonymization in the field of biomedical imaging, showcasing the capabilities of the model in addressing challenges related to data scarcity and privacy concerns. Our code and model weights are publicly accessible on our GitHub repository at https://github.com/mobaidoctor/med-ddpm/, facilitating reproducibility.

In recent years, self-supervised learning (SSL) frameworks have been extensively applied to sensor-based Human Activity Recognition (HAR) in order to learn deep representations without data annotations. While SSL frameworks reach performance almost comparable to supervised models, studies on interpreting representations learnt by SSL models are limited. Nevertheless, modern explainability methods could help to unravel the differences between SSL and supervised representations: how they are being learnt, what properties of input data they preserve, and when SSL can be chosen over supervised training. In this paper, we aim to analyze deep representations of two recent SSL frameworks, namely SimCLR and VICReg. Specifically, the emphasis is made on (i) comparing the robustness of supervised and SSL models to corruptions in input data; (ii) explaining predictions of deep learning models using saliency maps and highlighting what input channels are mostly used for predicting various activities; (iii) exploring properties encoded in SSL and supervised representations using probing. Extensive experiments on two single-device datasets (MobiAct and UCI-HAR) have shown that self-supervised learning representations are significantly more robust to noise in unseen data compared to supervised models. In contrast, features learnt by the supervised approaches are more homogeneous across subjects and better encode the nature of activities.

In this work, we present a new theoretical framework to measure re-identification risk in such user representations. Our framework, based on hypothesis testing, formally bounds the probability that an attacker may be able to obtain the identity of a user from their representation. As an application, we show how our framework is general enough to model important real-world applications such as the Chrome's Topics API for interest-based advertising. We complement our theoretical bounds by showing provably good attack algorithms for re-identification that we use to estimate the re-identification risk in the Topics API. We believe this work provides a rigorous and interpretable notion of re-identification risk and a framework to measure it that can be used to inform real-world applications.

Reed-Muller (RM) codes achieve the capacity of general binary-input memoryless symmetric channels and are conjectured to have a comparable performance to that of random codes in terms of scaling laws. However, such results are established assuming maximum-likelihood decoders for general code parameters. Also, RM codes only admit limited sets of rates. Efficient decoders such as successive cancellation list (SCL) decoder and recently-introduced recursive projection-aggregation (RPA) decoders are available for RM codes at finite lengths. In this paper, we focus on subcodes of RM codes with flexible rates. We first extend the RPA decoding algorithm to RM subcodes. To lower the complexity of our decoding algorithm, referred to as subRPA, we investigate different approaches to prune the projections. Next, we derive the soft-decision based version of our algorithm, called soft-subRPA, that not only improves upon the performance of subRPA but also enables a differentiable decoding algorithm. Building upon the soft-subRPA algorithm, we then provide a framework for training a machine learning (ML) model to search for \textit{good} sets of projections that minimize the decoding error rate. Training our ML model enables achieving very close to the performance of full-projection decoding with a significantly smaller number of projections. We also show that the choice of the projections in decoding RM subcodes matters significantly, and our ML-aided projection pruning scheme is able to find a \textit{good} selection, i.e., with negligible performance degradation compared to the full-projection case, given a reasonable number of projections.

Application for semantic segmentation models in areas such as autonomous vehicles and human computer interaction require real-time predictive capabilities. The challenges of addressing real-time application is amplified by the need to operate on resource constrained hardware. Whilst development of real-time methods for these platforms has increased, these models are unable to sufficiently reason about uncertainty present when applied on embedded real-time systems. This paper addresses this by combining deep feature extraction from pre-trained models with Bayesian regression and moment propagation for uncertainty aware predictions. We demonstrate how the proposed method can yield meaningful epistemic uncertainty on embedded hardware in real-time whilst maintaining predictive performance.

Predicting fund performance is beneficial to both investors and fund managers, and yet is a challenging task. In this paper, we have tested whether deep learning models can predict fund performance more accurately than traditional statistical techniques. Fund performance is typically evaluated by the Sharpe ratio, which represents the risk-adjusted performance to ensure meaningful comparability across funds. We calculated the annualised Sharpe ratios based on the monthly returns time series data for more than 600 open-end mutual funds investing in listed large-cap equities in the United States. We find that long short-term memory (LSTM) and gated recurrent units (GRUs) deep learning methods, both trained with modern Bayesian optimization, provide higher accuracy in forecasting funds' Sharpe ratios than traditional statistical ones. An ensemble method, which combines forecasts from LSTM and GRUs, achieves the best performance of all models. There is evidence to say that deep learning and ensembling offer promising solutions in addressing the challenge of fund performance forecasting.

A central quest of probing is to uncover how pre-trained models encode a linguistic property within their representations. An encoding, however, might be spurious-i.e., the model might not rely on it when making predictions. In this paper, we try to find encodings that the model actually uses, introducing a usage-based probing setup. We first choose a behavioral task which cannot be solved without using the linguistic property. Then, we attempt to remove the property by intervening on the model's representations. We contend that, if an encoding is used by the model, its removal should harm the performance on the chosen behavioral task. As a case study, we focus on how BERT encodes grammatical number, and on how it uses this encoding to solve the number agreement task. Experimentally, we find that BERT relies on a linear encoding of grammatical number to produce the correct behavioral output. We also find that BERT uses a separate encoding of grammatical number for nouns and verbs. Finally, we identify in which layers information about grammatical number is transferred from a noun to its head verb.