In typical multimodal contrastive learning, such as CLIP, encoders produce one point in the latent representation space for each input. However, one-point representation has difficulty in capturing the relationship and the similarity structure of a huge amount of instances in the real world. For richer classes of the similarity, we propose the use of weighted point clouds, namely, sets of pairs of weight and vector, as representations of instances. In this work, we theoretically show the benefit of our proposed method through a new understanding of the contrastive loss of CLIP, which we call symmetric InfoNCE. We clarify that the optimal similarity that minimizes symmetric InfoNCE is the pointwise mutual information, and show an upper bound of excess risk on downstream classification tasks of representations that achieve the optimal similarity. In addition, we show that our proposed similarity based on weighted point clouds consistently achieves the optimal similarity. To verify the effectiveness of our proposed method, we demonstrate pretraining of text-image representation models and classification tasks on common benchmarks.
In recent years, Event Sound Source Localization has been widely applied in various fields. Recent works typically relying on the contrastive learning framework show impressive performance. However, all work is based on large relatively simple datasets. It's also crucial to understand and analyze human behaviors (actions and interactions of people), voices, and sounds in chaotic events in many applications, e.g., crowd management, and emergency response services. In this paper, we apply the existing model to a more complex dataset, explore the influence of parameters on the model, and propose a semi-supervised improvement method SemiPL. With the increase in data quantity and the influence of label quality, self-supervised learning will be an unstoppable trend. The experiment shows that the parameter adjustment will positively affect the existing model. In particular, SSPL achieved an improvement of 12.2% cIoU and 0.56% AUC in Chaotic World compared to the results provided. The code is available at: https://github.com/ly245422/SSPL
Given the emergence of deep learning, digital pathology has gained popularity for cancer diagnosis based on histology images. Deep weakly supervised object localization (WSOL) models can be trained to classify histology images according to cancer grade and identify regions of interest (ROIs) for interpretation, using inexpensive global image-class annotations. A WSOL model initially trained on some labeled source image data can be adapted using unlabeled target data in cases of significant domain shifts caused by variations in staining, scanners, and cancer type. In this paper, we focus on source-free (unsupervised) domain adaptation (SFDA), a challenging problem where a pre-trained source model is adapted to a new target domain without using any source domain data for privacy and efficiency reasons. SFDA of WSOL models raises several challenges in histology, most notably because they are not intended to adapt for both classification and localization tasks. In this paper, 4 state-of-the-art SFDA methods, each one representative of a main SFDA family, are compared for WSOL in terms of classification and localization accuracy. They are the SFDA-Distribution Estimation, Source HypOthesis Transfer, Cross-Domain Contrastive Learning, and Adaptively Domain Statistics Alignment. Experimental results on the challenging Glas (smaller, breast cancer) and Camelyon16 (larger, colon cancer) histology datasets indicate that these SFDA methods typically perform poorly for localization after adaptation when optimized for classification.
Emotional Text-to-Speech (E-TTS) synthesis has garnered significant attention in recent years due to its potential to revolutionize human-computer interaction. However, current E-TTS approaches often struggle to capture the intricacies of human emotions, primarily relying on oversimplified emotional labels or single-modality input. In this paper, we introduce the Unified Multimodal Prompt-Induced Emotional Text-to-Speech System (UMETTS), a novel framework that leverages emotional cues from multiple modalities to generate highly expressive and emotionally resonant speech. The core of UMETTS consists of two key components: the Emotion Prompt Alignment Module (EP-Align) and the Emotion Embedding-Induced TTS Module (EMI-TTS). (1) EP-Align employs contrastive learning to align emotional features across text, audio, and visual modalities, ensuring a coherent fusion of multimodal information. (2) Subsequently, EMI-TTS integrates the aligned emotional embeddings with state-of-the-art TTS models to synthesize speech that accurately reflects the intended emotions. Extensive evaluations show that UMETTS achieves significant improvements in emotion accuracy and speech naturalness, outperforming traditional E-TTS methods on both objective and subjective metrics.
The representation of events in text plays a significant role in various NLP tasks. Recent research demonstrates that contrastive learning has the ability to improve event comprehension capabilities of Pre-trained Language Models (PLMs) and enhance the performance of event representation learning. However, the efficacy of event representation learning based on contrastive learning and PLMs is limited by the short length of event texts. The length of event texts differs significantly from the text length used in the pre-training of PLMs. As a result, there is inconsistency in the distribution of text length between pre-training and event representation learning, which may undermine the learning process of event representation based on PLMs. In this study, we present PromptCL, a novel framework for event representation learning that effectively elicits the capabilities of PLMs to comprehensively capture the semantics of short event texts. PromptCL utilizes a Prompt template borrowed from prompt learning to expand the input text during Contrastive Learning. This helps in enhancing the event representation learning by providing a structured outline of the event components. Moreover, we propose Subject-Predicate-Object (SPO) word order and Event-oriented Masked Language Modeling (EventMLM) to train PLMs to understand the relationships between event components. Our experimental results demonstrate that PromptCL outperforms state-of-the-art baselines on event related tasks. Additionally, we conduct a thorough analysis and demonstrate that using a prompt results in improved generalization capabilities for event representations. Our code will be available at https://github.com/YuboFeng2023/PromptCL.
Efficiently capturing consistent and complementary semantic features in a multimodal conversation context is crucial for Multimodal Emotion Recognition in Conversation (MERC). Existing methods mainly use graph structures to model dialogue context semantic dependencies and employ Graph Neural Networks (GNN) to capture multimodal semantic features for emotion recognition. However, these methods are limited by some inherent characteristics of GNN, such as over-smoothing and low-pass filtering, resulting in the inability to learn long-distance consistency information and complementary information efficiently. Since consistency and complementarity information correspond to low-frequency and high-frequency information, respectively, this paper revisits the problem of multimodal emotion recognition in conversation from the perspective of the graph spectrum. Specifically, we propose a Graph-Spectrum-based Multimodal Consistency and Complementary collaborative learning framework GS-MCC. First, GS-MCC uses a sliding window to construct a multimodal interaction graph to model conversational relationships and uses efficient Fourier graph operators to extract long-distance high-frequency and low-frequency information, respectively. Then, GS-MCC uses contrastive learning to construct self-supervised signals that reflect complementarity and consistent semantic collaboration with high and low-frequency signals, thereby improving the ability of high and low-frequency information to reflect real emotions. Finally, GS-MCC inputs the collaborative high and low-frequency information into the MLP network and softmax function for emotion prediction. Extensive experiments have proven the superiority of the GS-MCC architecture proposed in this paper on two benchmark data sets.
Taking inspiration from Set Theory, we introduce SetCSE, an innovative information retrieval framework. SetCSE employs sets to represent complex semantics and incorporates well-defined operations for structured information querying under the provided context. Within this framework, we introduce an inter-set contrastive learning objective to enhance comprehension of sentence embedding models concerning the given semantics. Furthermore, we present a suite of operations, including SetCSE intersection, difference, and operation series, that leverage sentence embeddings of the enhanced model for complex sentence retrieval tasks. Throughout this paper, we demonstrate that SetCSE adheres to the conventions of human language expressions regarding compounded semantics, provides a significant enhancement in the discriminatory capability of underlying sentence embedding models, and enables numerous information retrieval tasks involving convoluted and intricate prompts which cannot be achieved using existing querying methods.
Multi-modal entity alignment (MMEA) aims to identify equivalent entity pairs across different multi-modal knowledge graphs (MMKGs). Existing approaches focus on how to better encode and aggregate information from different modalities. However, it is not trivial to leverage multi-modal knowledge in entity alignment due to the modal heterogeneity. In this paper, we propose a Multi-Grained Interaction framework for Multi-Modal Entity Alignment (MIMEA), which effectively realizes multi-granular interaction within the same modality or between different modalities. MIMEA is composed of four modules: i) a Multi-modal Knowledge Embedding module, which extracts modality-specific representations with multiple individual encoders; ii) a Probability-guided Modal Fusion module, which employs a probability guided approach to integrate uni-modal representations into joint-modal embeddings, while considering the interaction between uni-modal representations; iii) an Optimal Transport Modal Alignment module, which introduces an optimal transport mechanism to encourage the interaction between uni-modal and joint-modal embeddings; iv) a Modal-adaptive Contrastive Learning module, which distinguishes the embeddings of equivalent entities from those of non-equivalent ones, for each modality. Extensive experiments conducted on two real-world datasets demonstrate the strong performance of MIMEA compared to the SoTA. Datasets and code have been submitted as supplementary materials.
Document layout analysis involves understanding the arrangement of elements within a document. This paper navigates the complexities of understanding various elements within document images, such as text, images, tables, and headings. The approach employs an advanced Transformer-based object detection network as an innovative graphical page object detector for identifying tables, figures, and displayed elements. We introduce a query encoding mechanism to provide high-quality object queries for contrastive learning, enhancing efficiency in the decoder phase. We also present a hybrid matching scheme that integrates the decoder's original one-to-one matching strategy with the one-to-many matching strategy during the training phase. This approach aims to improve the model's accuracy and versatility in detecting various graphical elements on a page. Our experiments on PubLayNet, DocLayNet, and PubTables benchmarks show that our approach outperforms current state-of-the-art methods. It achieves an average precision of 97.3% on PubLayNet, 81.6% on DocLayNet, and 98.6 on PubTables, demonstrating its superior performance in layout analysis. These advancements not only enhance the conversion of document images into editable and accessible formats but also streamline information retrieval and data extraction processes.
Image-text matching remains a challenging task due to heterogeneous semantic diversity across modalities and insufficient distance separability within triplets. Different from previous approaches focusing on enhancing multi-modal representations or exploiting cross-modal correspondence for more accurate retrieval, in this paper we aim to leverage the knowledge transfer between peer branches in a boosting manner to seek a more powerful matching model. Specifically, we propose a brand-new Deep Boosting Learning (DBL) algorithm, where an anchor branch is first trained to provide insights into the data properties, with a target branch gaining more advanced knowledge to develop optimal features and distance metrics. Concretely, an anchor branch initially learns the absolute or relative distance between positive and negative pairs, providing a foundational understanding of the particular network and data distribution. Building upon this knowledge, a target branch is concurrently tasked with more adaptive margin constraints to further enlarge the relative distance between matched and unmatched samples. Extensive experiments validate that our DBL can achieve impressive and consistent improvements based on various recent state-of-the-art models in the image-text matching field, and outperform related popular cooperative strategies, e.g., Conventional Distillation, Mutual Learning, and Contrastive Learning. Beyond the above, we confirm that DBL can be seamlessly integrated into their training scenarios and achieve superior performance under the same computational costs, demonstrating the flexibility and broad applicability of our proposed method. Our code is publicly available at: https://github.com/Paranioar/DBL.
CLIP showcases exceptional cross-modal matching capabilities due to its training on image-text contrastive learning tasks. However, without specific optimization for unimodal scenarios, its performance in single-modality feature extraction might be suboptimal. Despite this, some studies have directly used CLIP's image encoder for tasks like few-shot classification, introducing a misalignment between its pre-training objectives and feature extraction methods. This inconsistency can diminish the quality of the image's feature representation, adversely affecting CLIP's effectiveness in target tasks. In this paper, we view text features as precise neighbors of image features in CLIP's space and present a novel CrOss-moDal nEighbor Representation(CODER) based on the distance structure between images and their neighbor texts. This feature extraction method aligns better with CLIP's pre-training objectives, thereby fully leveraging CLIP's robust cross-modal capabilities. The key to construct a high-quality CODER lies in how to create a vast amount of high-quality and diverse texts to match with images. We introduce the Auto Text Generator(ATG) to automatically generate the required texts in a data-free and training-free manner. We apply CODER to CLIP's zero-shot and few-shot image classification tasks. Experiment results across various datasets and models confirm CODER's effectiveness. Code is available at:https://github.com/YCaigogogo/CVPR24-CODER.
Click-through rate (CTR) prediction is crucial for personalized online services. Sample-level retrieval-based models, such as RIM, have demonstrated remarkable performance. However, they face challenges including inference inefficiency and high resource consumption due to the retrieval process, which hinder their practical application in industrial settings. To address this, we propose a universal plug-and-play \underline{r}etrieval-\underline{o}riented \underline{k}nowledge (\textbf{\name}) framework that bypasses the real retrieval process. The framework features a knowledge base that preserves and imitates the retrieved \& aggregated representations using a decomposition-reconstruction paradigm. Knowledge distillation and contrastive learning optimize the knowledge base, enabling the integration of retrieval-enhanced representations with various CTR models. Experiments on three large-scale datasets demonstrate \name's exceptional compatibility and performance, with the neural knowledge base serving as an effective surrogate for the retrieval pool. \name surpasses the teacher model while maintaining superior inference efficiency and demonstrates the feasibility of distilling knowledge from non-parametric methods using a parametric approach. These results highlight \name's strong potential for real-world applications and its ability to transform retrieval-based methods into practical solutions. Our implementation code is available to support reproducibility in \url{https://github.com/HSLiu-Initial/ROK.git}.
Sequential recommendation is one of the important branches of recommender system, aiming to achieve personalized recommended items for the future through the analysis and prediction of users' ordered historical interactive behaviors. However, along with the growth of the user volume and the increasingly rich behavioral information, how to understand and disentangle the user's interactive multi-intention effectively also poses challenges to behavior prediction and sequential recommendation. In light of these challenges, we propose a Contrastive Learning sequential recommendation method based on Multi-Intention Disentanglement (MIDCL). In our work, intentions are recognized as dynamic and diverse, and user behaviors are often driven by current multi-intentions, which means that the model needs to not only mine the most relevant implicit intention for each user, but also impair the influence from irrelevant intentions. Therefore, we choose Variational Auto-Encoder (VAE) to realize the disentanglement of users' multi-intentions. We propose two types of contrastive learning paradigms for finding the most relevant user's interactive intention, and maximizing the mutual information of positive sample pairs, respectively. Experimental results show that MIDCL not only has significant superiority over most existing baseline methods, but also brings a more interpretable case to the research about intention-based prediction and recommendation.
Heterogeneity resulting from label distribution skew and data scarcity can lead to inaccuracy and unfairness in intelligent communication applications that mainly rely on distributed computing. To deal with it, this paper proposes a novel personalized federated learning algorithm, named Federated Contrastive Shareable Representations (FedCoSR), to facilitate knowledge sharing among clients while maintaining data privacy. Specifically, parameters of local models' shallow layers and typical local representations are both considered shareable information for the server and aggregated globally. To address poor performance caused by label distribution skew among clients, contrastive learning is adopted between local and global representations to enrich local knowledge. Additionally, to ensure fairness for clients with scarce data, FedCoSR introduces adaptive local aggregation to coordinate the global model involvement in each client. Our simulations demonstrate FedCoSR's effectiveness in mitigating label heterogeneity by achieving accuracy and fairness improvements over existing methods on datasets with varying degrees of label heterogeneity.
Understanding the severity of conditions shown in images in medical diagnosis is crucial, serving as a key guide for clinical assessment, treatment, as well as evaluating longitudinal progression. This paper proposes Con- PrO: a novel representation learning method for severity assessment in medical images using Contrastive learningintegrated Preference Optimization. Different from conventional contrastive learning methods that maximize the distance between classes, ConPrO injects into the latent vector the distance preference knowledge between various severity classes and the normal class. We systematically examine the key components of our framework to illuminate how contrastive prediction tasks acquire valuable representations. We show that our representation learning framework offers valuable severity ordering in the feature space while outperforming previous state-of-the-art methods on classification tasks. We achieve a 6% and 20% relative improvement compared to a supervised and a self-supervised baseline, respectively. In addition, we derived discussions on severity indicators and related applications of preference comparison in the medical domain.
The classification of sleep stages is a pivotal aspect of diagnosing sleep disorders and evaluating sleep quality. However, the conventional manual scoring process, conducted by clinicians, is time-consuming and prone to human bias. Recent advancements in deep learning have substantially propelled the automation of sleep stage classification. Nevertheless, challenges persist, including the need for large datasets with labels and the inherent biases in human-generated annotations. This paper introduces NeuroNet, a self-supervised learning (SSL) framework designed to effectively harness unlabeled single-channel sleep electroencephalogram (EEG) signals by integrating contrastive learning tasks and masked prediction tasks. NeuroNet demonstrates superior performance over existing SSL methodologies through extensive experimentation conducted across three polysomnography (PSG) datasets. Additionally, this study proposes a Mamba-based temporal context module to capture the relationships among diverse EEG epochs. Combining NeuroNet with the Mamba-based temporal context module has demonstrated the capability to achieve, or even surpass, the performance of the latest supervised learning methodologies, even with a limited amount of labeled data. This study is expected to establish a new benchmark in sleep stage classification, promising to guide future research and applications in the field of sleep analysis.
Few-shot Named Entity Recognition (NER) aims to extract named entities using only a limited number of labeled examples. Existing contrastive learning methods often suffer from insufficient distinguishability in context vector representation because they either solely rely on label semantics or completely disregard them. To tackle this issue, we propose a unified label-aware token-level contrastive learning framework. Our approach enriches the context by utilizing label semantics as suffix prompts. Additionally, it simultaneously optimizes context-context and context-label contrastive learning objectives to enhance generalized discriminative contextual representations.Extensive experiments on various traditional test domains (OntoNotes, CoNLL'03, WNUT'17, GUM, I2B2) and the large-scale few-shot NER dataset (FEWNERD) demonstrate the effectiveness of our approach. It outperforms prior state-of-the-art models by a significant margin, achieving an average absolute gain of 7% in micro F1 scores across most scenarios. Further analysis reveals that our model benefits from its powerful transfer capability and improved contextual representations.
We present COCOLA (Coherence-Oriented Contrastive Learning for Audio), a contrastive learning method for musical audio representations that captures the harmonic and rhythmic coherence between samples. Our method operates at the level of the stems composing music tracks and can input features obtained via Harmonic-Percussive Separation (HPS). COCOLA allows the objective evaluation of generative models for music accompaniment generation, which are difficult to benchmark with established metrics. In this regard, we evaluate recent music accompaniment generation models, demonstrating the effectiveness of the proposed method. We release the model checkpoints trained on public datasets containing separate stems (MUSDB18-HQ, MoisesDB, Slakh2100, and CocoChorales).
Contrastive learning is a model pre-training technique by first creating similar views of the original data, and then encouraging the data and its corresponding views to be close in the embedding space. Contrastive learning has witnessed success in image and natural language data, thanks to the domain-specific augmentation techniques that are both intuitive and effective. Nonetheless, in tabular domain, the predominant augmentation technique for creating views is through corrupting tabular entries via swapping values, which is not as sound or effective. We propose a simple yet powerful improvement to this augmentation technique: corrupting tabular data conditioned on class identity. Specifically, when corrupting a specific tabular entry from an anchor row, instead of randomly sampling a value in the same feature column from the entire table uniformly, we only sample from rows that are identified to be within the same class as the anchor row. We assume the semi-supervised learning setting, and adopt the pseudo labeling technique for obtaining class identities over all table rows. We also explore the novel idea of selecting features to be corrupted based on feature correlation structures. Extensive experiments show that the proposed approach consistently outperforms the conventional corruption method for tabular data classification tasks. Our code is available at https://github.com/willtop/Tabular-Class-Conditioned-SSL.
Named entity recognition (NER) is a fundamental task in natural language processing that involves identifying and classifying entities in sentences into pre-defined types. It plays a crucial role in various research fields, including entity linking, question answering, and online product recommendation. Recent studies have shown that incorporating multilingual and multimodal datasets can enhance the effectiveness of NER. This is due to language transfer learning and the presence of shared implicit features across different modalities. However, the lack of a dataset that combines multilingualism and multimodality has hindered research exploring the combination of these two aspects, as multimodality can help NER in multiple languages simultaneously. In this paper, we aim to address a more challenging task: multilingual and multimodal named entity recognition (MMNER), considering its potential value and influence. Specifically, we construct a large-scale MMNER dataset with four languages (English, French, German and Spanish) and two modalities (text and image). To tackle this challenging MMNER task on the dataset, we introduce a new model called 2M-NER, which aligns the text and image representations using contrastive learning and integrates a multimodal collaboration module to effectively depict the interactions between the two modalities. Extensive experimental results demonstrate that our model achieves the highest F1 score in multilingual and multimodal NER tasks compared to some comparative and representative baselines. Additionally, in a challenging analysis, we discovered that sentence-level alignment interferes a lot with NER models, indicating the higher level of difficulty in our dataset.