Sequential abstractive neural summarizers often do not use the underlying structure in the input article or dependencies between the input sentences. This structure is essential to integrate and consolidate information from different parts of the text. To address this shortcoming, we propose a hierarchy-aware graph neural network (HierGNN) which captures such dependencies through three main steps: 1) learning a hierarchical document structure through a latent structure tree learned by a sparse matrix-tree computation; 2) propagating sentence information over this structure using a novel message-passing node propagation mechanism to identify salient information; 3) using graph-level attention to concentrate the decoder on salient information. Experiments confirm HierGNN improves strong sequence models such as BART, with a 0.55 and 0.75 margin in average ROUGE-1/2/L for CNN/DM and XSum. Further human evaluation demonstrates that summaries produced by our model are more relevant and less redundant than the baselines, into which HierGNN is incorporated. We also find HierGNN synthesizes summaries by fusing multiple source sentences more, rather than compressing a single source sentence, and that it processes long inputs more effectively.
Speech is the fundamental means of communication between humans. The advent of AI and sophisticated speech technologies have led to the rapid proliferation of human-to-computer-based interactions, fueled primarily by Automatic Speech Recognition (ASR) systems. ASR systems normally take human speech in the form of audio and convert it into words, but for some users, it cannot decode the speech, and any output text is filled with errors that are incomprehensible to the human reader. These systems do not work equally for everyone and actually hinder the productivity of some users. In this paper, we present research that addresses ASR biases against gender, race, and the sick and disabled, while exploring studies that propose ASR debiasing techniques for mitigating these discriminations. We also discuss techniques for designing a more accessible and inclusive ASR technology. For each approach surveyed, we also provide a summary of the investigation and methods applied, the ASR systems and corpora used, and the research findings, and highlight their strengths and/or weaknesses. Finally, we propose future opportunities for Natural Language Processing researchers to explore in the next level creation of ASR technologies.
In this paper, we propose two novel augmentation methods 1) audio-language MixGen (AL-MixGen) and 2) multi-level test-time augmentation (Multi-TTA) for audio-language learning. Inspired by MixGen, which is originally applied to vision-language learning, we introduce an augmentation method for the audio-language domain. We also explore the impact of test-time augmentations and present Multi-TTA which generalizes test-time augmentation over multiple layers of a deep learning model. Incorporating AL-MixGen and Multi-TTA into the baseline achieves 47.5 SPIDEr on audio captioning, which is an +18.2% over the baseline and outperforms the state-of-the-art approach with a 5x smaller model. In audio-text retrieval, the proposed methods surpass the baseline performance as well.
Embedding based product recommendations have gained popularity in recent years due to its ability to easily integrate to large-scale systems and allowing nearest neighbor searches in real-time. The bulk of studies in this area has predominantly been focused on similar item recommendations. Research on complementary item recommendations, on the other hand, still remains considerably under-explored. We define similar items as items that are interchangeable in terms of their utility and complementary items as items that serve different purposes, yet are compatible when used with one another. In this paper, we apply a novel approach to finding complementary items by leveraging dual embedding representations for products. We demonstrate that the notion of relatedness discovered in NLP for skip-gram negative sampling (SGNS) models translates effectively to the concept of complementarity when training item representations using co-purchase data. Since sparsity of purchase data is a major challenge in real-world scenarios, we further augment the model using synthetic samples to extend coverage. This allows the model to provide complementary recommendations for items that do not share co-purchase data by leveraging other abundantly available data modalities such as images, text, clicks etc. We establish the effectiveness of our approach in improving both coverage and quality of recommendations on real world data for a major online retail company. We further show the importance of task specific hyperparameter tuning in training SGNS. Our model is effective yet simple to implement, making it a great candidate for generating complementary item recommendations at any e-commerce website.
With the availability of voice-enabled devices such as smart phones, mental health disorders could be detected and treated earlier, particularly post-pandemic. The current methods involve extracting features directly from audio signals. In this paper, two methods are used to enrich voice analysis for depression detection: graph transformation of voice signals, and natural language processing of the transcript based on representational learning, fused together to produce final class labels. The results of experiments with the DAIC-WOZ dataset suggest that integration of text-based voice classification and learning from low level and graph-based voice signal features can improve the detection of mental disorders like depression.
Most existing vision-language pre-training (VLP) approaches adopt cross-modal masked language modeling (CMLM) to learn vision-language associations. However, we find that CMLM is insufficient for this purpose according to our observations: (1) Modality bias: a considerable amount of masked tokens in CMLM can be recovered with only the language information, ignoring the visual inputs. (2) Under-utilization of the unmasked tokens: CMLM primarily focuses on the masked token but it cannot simultaneously leverage other tokens to learn vision-language associations. To handle those limitations, we propose EPIC (lEveraging Per Image-Token Consistency for vision-language pre-training). In EPIC, for each image-sentence pair, we mask tokens that are salient to the image (i.e., Saliency-based Masking Strategy) and replace them with alternatives sampled from a language model (i.e., Inconsistent Token Generation Procedure), and then the model is required to determine for each token in the sentence whether they are consistent with the image (i.e., Image-Text Consistent Task). The proposed EPIC method is easily combined with pre-training methods. Extensive experiments show that the combination of the EPIC method and state-of-the-art pre-training approaches, including ViLT, ALBEF, METER, and X-VLM, leads to significant improvements on downstream tasks.
Deep learning usually relies on training large-scale data samples to achieve better performance. However, over-fitting based on training data always remains a problem. Scholars have proposed various strategies, such as feature dropping and feature mixing, to improve the generalization continuously. For the same purpose, we subversively propose a novel training method, Feature Weaken, which can be regarded as a data augmentation method. Feature Weaken constructs the vicinal data distribution with the same cosine similarity for model training by weakening features of the original samples. In especially, Feature Weaken changes the spatial distribution of samples, adjusts sample boundaries, and reduces the gradient optimization value of back-propagation. This work can not only improve the classification performance and generalization of the model, but also stabilize the model training and accelerate the model convergence. We conduct extensive experiments on classical deep convolution neural models with five common image classification datasets and the Bert model with four common text classification datasets. Compared with the classical models or the generalization improvement methods, such as Dropout, Mixup, Cutout, and CutMix, Feature Weaken shows good compatibility and performance. We also use adversarial samples to perform the robustness experiments, and the results show that Feature Weaken is effective in improving the robustness of the model.
Unsupervised generation of 3D-aware clothed humans with various appearances and controllable geometries is important for creating virtual human avatars and other AR/VR applications. Existing methods are either limited to rigid object modeling, or not generative and thus unable to generate high-quality virtual humans and animate them. In this work, we propose AvatarGen, the first method that enables not only geometry-aware clothed human synthesis with high-fidelity appearances but also disentangled human animation controllability, while only requiring 2D images for training. Specifically, we decompose the generative 3D human synthesis into pose-guided mapping and canonical representation with predefined human pose and shape, such that the canonical representation can be explicitly driven to different poses and shapes with the guidance of a 3D parametric human model SMPL. AvatarGen further introduces a deformation network to learn non-rigid deformations for modeling fine-grained geometric details and pose-dependent dynamics. To improve the geometry quality of the generated human avatars, it leverages the signed distance field as geometric proxy, which allows more direct regularization from the 3D geometric priors of SMPL. Benefiting from these designs, our method can generate animatable 3D human avatars with high-quality appearance and geometry modeling, significantly outperforming previous 3D GANs. Furthermore, it is competent for many applications, e.g., single-view reconstruction, re-animation, and text-guided synthesis/editing. Code and pre-trained model will be available at http://jeff95.me/projects/avatargen.html.
Text classification is one of the fundamental tasks in natural language processing to label an open-ended text and is useful for various applications such as sentiment analysis. In this paper, we discuss various classification approaches for Khmer text, ranging from a classical TF-IDF algorithm with support vector machine classifier to modern word embedding-based neural network classifiers including linear layer model, recurrent neural network and convolutional neural network. A Khmer word embedding model is trained on a 30-million-Khmer-word corpus to construct word vector representations that are used to train three different neural network classifiers. We evaluate the performance of different approaches on a news article dataset for both multi-class and multi-label text classification tasks. The result suggests that neural network classifiers using a word embedding model consistently outperform the traditional classifier using TF-IDF. The recurrent neural network classifier provides a slightly better result compared to the convolutional network and the linear layer network.
Text detection/localization, as an important task in computer vision, has witnessed substantialadvancements in methodology and performance with convolutional neural networks. However, the vastmajority of popular methods use rectangles or quadrangles to describe text regions. These representationshave inherent drawbacks, especially relating to dense adjacent text and loose regional text boundaries,which usually cause difficulty detecting arbitrarily shaped text. In this paper, we propose a novel text regionrepresentation method, with a robust pipeline, which can precisely detect dense adjacent text instances witharbitrary shapes. We consider a text instance to be composed of an adaptive central text region mask anda corresponding expanding ratio between the central text region and the full text region. More specifically,our pipeline generates adaptive central text regions and corresponding expanding ratios with a proposedtraining strategy, followed by a new proposed post-processing algorithm which expands central text regionsto the complete text instance with the corresponding expanding ratios. We demonstrated that our new textregion representation is effective, and that the pipeline can precisely detect closely adjacent text instances ofarbitrary shapes. Experimental results on common datasets demonstrate superior performance o