Data augmentation (DA) is indispensable in modern machine learning and deep neural networks. The basic idea of DA is to construct new training data to improve the model's generalization by adding slightly disturbed versions of existing data or synthesizing new data. In this work, we review a small but essential subset of DA -- Mix-based Data Augmentation (MixDA) that generates novel samples by mixing multiple examples. Unlike conventional DA approaches based on a single-sample operation or requiring domain knowledge, MixDA is more general in creating a broad spectrum of new data and has received increasing attention in the community. We begin with proposing a new taxonomy classifying MixDA into, Mixup-based, Cutmix-based, and hybrid approaches according to a hierarchical view of the data mix. Various MixDA techniques are then comprehensively reviewed in a more fine-grained way. Owing to its generalization, MixDA has penetrated a variety of applications which are also completely reviewed in this work. We also examine why MixDA works from different aspects of improving model performance, generalization, and calibration while explaining the model behavior based on the properties of MixDA. Finally, we recapitulate the critical findings and fundamental challenges of current MixDA studies, and outline the potential directions for future works. Different from previous related works that summarize the DA approaches in a specific domain (e.g., images or natural language processing) or only review a part of MixDA studies, we are the first to provide a systematical survey of MixDA in terms of its taxonomy, methodology, applications, and explainability. This work can serve as a roadmap to MixDA techniques and application reviews while providing promising directions for researchers interested in this exciting area.

To improve the performance of the dual-encoder retriever, one effective approach is knowledge distillation from the cross-encoder ranker. Existing works construct the candidate passages following the supervised learning setting where a query is paired with a positive passage and a batch of negatives. However, through empirical observation, we find that even the hard negatives from advanced methods are still too trivial for the teacher to distinguish, preventing the teacher from transferring abundant dark knowledge to the student through its soft label. To alleviate this issue, we propose ADAM, a knowledge distillation framework that can better transfer the dark knowledge held in the teacher with Adaptive Dark exAMples. Different from previous works that only rely on one positive and hard negatives as candidate passages, we create dark examples that all have moderate relevance to the query through mixing-up and masking in discrete space. Furthermore, as the quality of knowledge held in different training instances varies as measured by the teacher's confidence score, we propose a self-paced distillation strategy that adaptively concentrates on a subset of high-quality instances to conduct our dark-example-based knowledge distillation to help the student learn better. We conduct experiments on two widely-used benchmarks and verify the effectiveness of our method.

We present Hybrid Infused Reranking for Passages Retrieval (HYRR), a framework for training rerankers based on a hybrid of BM25 and neural retrieval models. Retrievers based on hybrid models have been shown to outperform both BM25 and neural models alone. Our approach exploits this improved performance when training a reranker, leading to a robust reranking model. The reranker, a cross-attention neural model, is shown to be robust to different first-stage retrieval systems, achieving better performance than rerankers simply trained upon the first-stage retrievers in the multi-stage systems. We present evaluations on a supervised passage retrieval task using MS MARCO and zero-shot retrieval tasks using BEIR. The empirical results show strong performance on both evaluations.

Pre-trained language models (LMs) store knowledge in their parameters and can generate informative responses when used in conversational systems. However, LMs suffer from the problem of "hallucination:" they may generate plausible-looking statements that are irrelevant or factually incorrect. To address this problem, we propose a contrastive learning scheme, named MixCL. A novel mixed contrastive objective is proposed to explicitly optimize the implicit knowledge elicitation process of LMs, and thus reduce their hallucination in conversations. We also examine negative sampling strategies of retrieved hard negatives and model-generated negatives. We conduct experiments on Wizard-of-Wikipedia, a public, open-domain knowledge-grounded dialogue benchmark, and assess the effectiveness of MixCL. MixCL effectively reduces the hallucination of LMs in conversations and achieves the highest performance among LM-based dialogue agents in terms of relevancy and factuality. We show that MixCL achieves comparable performance to state-of-the-art KB-based approaches while enjoying notable advantages in terms of efficiency and scalability.

How you can train a model to learn and predict unseen data? The reason why machine learning models in general are becoming smarter is due to their dependency on using labeled data to help them discern between two similar objects. However, without these labeled datasets, you will encounter major obstacles when creating the most effective and trustworthy machine-learning model. Deep learning has been widely used to solve tasks such as Computer vision using supervised learning. However, as with many things in life, it comes with restrictions.

Your business may generate mountains of data. But are you taking the edge of the visions it would reveal? You can use machine learning, a branch of AI, to analyze your data and predicts future outcome or identify hidden patterns. Today, I'll cover two approaches named supervised and unsupervised machine learning. The significant difference between the two is how the training data is labeled.

We propose P4E, an identify-and-localize event detection framework that integrates the best of few-shot prompting and structured prediction. Our framework decomposes event detection into an identification task and a localization task. For the identification task, which we formulate as multi-label classification, we leverage cloze-based prompting to align our objective with the pre-training task of language models, allowing our model to quickly adapt to new event types. We then employ an event type-agnostic sequence labeling model to localize the event trigger conditioned on the identification output. This heterogeneous model design allows P4E to quickly learn new event types without sacrificing the ability to make structured predictions. Our experiments demonstrate the effectiveness of our proposed design, and P4E shows superior performance for few-shot event detection on benchmark datasets FewEvent and MAVEN and comparable performance to SOTA for fully-supervised event detection on ACE.

The robustness of machine learning algorithms to distributions shift is primarily discussed in the context of supervised learning (SL). As such, there is a lack of insight on the robustness of the representations learned from unsupervised methods, such as self-supervised learning (SSL) and auto-encoder based algorithms (AE), to distribution shift. We posit that the input-driven objectives of unsupervised algorithms lead to representations that are more robust to distribution shift than the target-driven objective of SL. We verify this by extensively evaluating the performance of SSL and AE on both synthetic and realistic distribution shift datasets. Following observations that the linear layer used for classification itself can be susceptible to spurious correlations, we evaluate the representations using a linear head trained on a small amount of out-of-distribution (OOD) data, to isolate the robustness of the learned representations from that of the linear head. We also develop "controllable" versions of existing realistic domain generalisation datasets with adjustable degrees of distribution shifts. This allows us to study the robustness of different learning algorithms under versatile yet realistic distribution shift conditions. Our experiments show that representations learned from unsupervised learning algorithms generalise better than SL under a wide variety of extreme as well as realistic distribution shifts.

Unsupervised domain adaptation reduces the reliance on data annotation in deep learning by adapting knowledge from a source to a target domain. For privacy and efficiency concerns, source-free domain adaptation extends unsupervised domain adaptation by adapting a pre-trained source model to an unlabeled target domain without accessing the source data. However, most existing source-free domain adaptation methods to date focus on the transductive setting, where the target training set is also the testing set. In this paper, we address source-free domain adaptation in the more realistic inductive setting, where the target training and testing sets are mutually exclusive. We propose a new semi-supervised fine-tuning method named Dual Moving Average Pseudo-Labeling (DMAPL) for source-free inductive domain adaptation. We first split the unlabeled training set in the target domain into a pseudo-labeled confident subset and an unlabeled less-confident subset according to the prediction confidence scores from the pre-trained source model. Then we propose a soft-label moving-average updating strategy for the unlabeled subset based on a moving-average prototypical classifier, which gradually adapts the source model towards the target domain. Experiments show that our proposed method achieves state-of-the-art performance and outperforms previous methods by large margins.

Contrastive learning (CL), which can extract the information shared between different contrastive views, has become a popular paradigm for vision representation learning. Inspired by the success in computer vision, recent work introduces CL into graph modeling, dubbed as graph contrastive learning (GCL). However, generating contrastive views in graphs is more challenging than that in images, since we have little prior knowledge on how to significantly augment a graph without changing its labels. We argue that typical data augmentation techniques (e.g., edge dropping) in GCL cannot generate diverse enough contrastive views to filter out noises. Moreover, previous GCL methods employ two view encoders with exactly the same neural architecture and tied parameters, which further harms the diversity of augmented views. To address this limitation, we propose a novel paradigm named model augmented GCL (MA-GCL), which will focus on manipulating the architectures of view encoders instead of perturbing graph inputs. Specifically, we present three easy-to-implement model augmentation tricks for GCL, namely asymmetric, random and shuffling, which can respectively help alleviate high- frequency noises, enrich training instances and bring safer augmentations. All three tricks are compatible with typical data augmentations. Experimental results show that MA-GCL can achieve state-of-the-art performance on node classification benchmarks by applying the three tricks on a simple base model. Extensive studies also validate our motivation and the effectiveness of each trick. (Code, data and appendix are available at https://github.com/GXM1141/MA-GCL. )