Different people have different facial expressions while speaking emotionally. A realistic facial animation system should consider such identity-specific speaking styles and facial idiosyncrasies to achieve high-degree of naturalness and plausibility. Existing approaches to personalized speech-driven 3D facial animation either use one-hot identity labels or rely-on person specific models which limit their scalability. We present a personalized speech-driven expressive 3D facial animation synthesis framework that models identity specific facial motion as latent representations (called as styles), and synthesizes novel animations given a speech input with the target style for various emotion categories. Our framework is trained in an end-to-end fashion and has a non-autoregressive encoder-decoder architecture with three main components: expression encoder, speech encoder and expression decoder. Since, expressive facial motion includes both identity-specific style and speech-related content information; expression encoder first disentangles facial motion sequences into style and content representations, respectively. Then, both of the speech encoder and the expression decoders input the extracted style information to update transformer layer weights during training phase. Our speech encoder also extracts speech phoneme label and duration information to achieve better synchrony within the non-autoregressive synthesis mechanism more effectively. Through detailed experiments, we demonstrate that our approach produces temporally coherent facial expressions from input speech while preserving the speaking styles of the target identities.
The pre-training architectures of large language models encompass various types, including autoencoding models, autoregressive models, and encoder-decoder models. We posit that any modality can potentially benefit from a large language model, as long as it undergoes vector quantization to become discrete tokens. Inspired by GLM, we propose a General Point Model (GPM) which seamlessly integrates autoencoding and autoregressive tasks in point cloud transformer. This model is versatile, allowing fine-tuning for downstream point cloud representation tasks, as well as unconditional and conditional generation tasks. GPM enhances masked prediction in autoencoding through various forms of mask padding tasks, leading to improved performance in point cloud understanding. Additionally, GPM demonstrates highly competitive results in unconditional point cloud generation tasks, even exhibiting the potential for conditional generation tasks by modifying the input's conditional information. Compared to models like Point-BERT, MaskPoint and PointMAE, our GPM achieves superior performance in point cloud understanding tasks. Furthermore, the integration of autoregressive and autoencoding within the same transformer underscores its versatility across different downstream tasks.
Machine learning is a powerful tool for extracting valuable information and making various predictions from diverse datasets. Traditional machine learning algorithms rely on well-defined input and output variables; however, there are scenarios where the separation between the input and output variables and the underlying, associated input and output layers of the model are unknown. Feature Selection (FS) and Neural Architecture Search (NAS) have emerged as promising solutions in such scenarios. This paper proposes MICS-EFS, a Model Input-Output Configuration Search with Embedded Feature Selection. The methodology explores internal dependencies in the complete input parameter space for classification tasks involving both 1D sensor time-series and 2D image data. MICS-EFS employs a modified encoder-decoder model and the Sequential Forward Search (SFS) algorithm, combining input-output configuration search with embedded feature selection. Experimental results demonstrate the superior performance of MICS-EFS compared to other FS algorithms. Across all tested datasets, MICS-EFS delivered an average accuracy improvement of 1.5% over baseline models, with the accuracy gains ranging from 0.5% to 5.9%. Moreover, the algorithm reduced feature dimensionality to just 2-5% of the original data, significantly enhancing computational efficiency. These results highlight the potential of MICS-EFS to improve model accuracy and efficiency in various machine learning tasks. Furthermore, the proposed method has been validated in a real-world industrial application focused on machining processes, underscoring its effectiveness and practicality in addressing complex input-output challenges.
Spatio-temporal graph neural networks (STGNNs) have gained popularity as a powerful tool for effectively modeling spatio-temporal dependencies in diverse real-world urban applications, including intelligent transportation and public safety. However, the black-box nature of STGNNs limits their interpretability, hindering their application in scenarios related to urban resource allocation and policy formulation. To bridge this gap, we propose an Explainable Spatio-Temporal Graph Neural Networks (STExplainer) framework that enhances STGNNs with inherent explainability, enabling them to provide accurate predictions and faithful explanations simultaneously. Our framework integrates a unified spatio-temporal graph attention network with a positional information fusion layer as the STG encoder and decoder, respectively. Furthermore, we propose a structure distillation approach based on the Graph Information Bottleneck (GIB) principle with an explainable objective, which is instantiated by the STG encoder and decoder. Through extensive experiments, we demonstrate that our STExplainer outperforms state-of-the-art baselines in terms of predictive accuracy and explainability metrics (i.e., sparsity and fidelity) on traffic and crime prediction tasks. Furthermore, our model exhibits superior representation ability in alleviating data missing and sparsity issues. The implementation code is available at: https://github.com/HKUDS/STExplainer.
Key information extraction (KIE) from scanned documents has gained increasing attention because of its applications in various domains. Although promising results have been achieved by some recent KIE approaches, they are usually built based on discriminative models, which lack the ability to handle optical character recognition (OCR) errors and require laborious token-level labelling. In this paper, we propose a novel generative end-to-end model, named GenKIE, to address the KIE task. GenKIE is a sequence-to-sequence multimodal generative model that utilizes multimodal encoders to embed visual, layout and textual features and a decoder to generate the desired output. Well-designed prompts are leveraged to incorporate the label semantics as the weakly supervised signals and entice the generation of the key information. One notable advantage of the generative model is that it enables automatic correction of OCR errors. Besides, token-level granular annotation is not required. Extensive experiments on multiple public real-world datasets show that GenKIE effectively generalizes over different types of documents and achieves state-of-the-art results. Our experiments also validate the model's robustness against OCR errors, making GenKIE highly applicable in real-world scenarios.
The assumption of conditional independence among observed variables, primarily used in the Variational Autoencoder (VAE) decoder modeling, has limitations when dealing with high-dimensional datasets or complex correlation structures among observed variables. To address this issue, we introduced the Cramer-Wold distance regularization, which can be computed in a closed-form, to facilitate joint distributional learning for high-dimensional datasets. Additionally, we introduced a two-step learning method to enable flexible prior modeling and improve the alignment between the aggregated posterior and the prior distribution. Furthermore, we provide theoretical distinctions from existing methods within this category. To evaluate the synthetic data generation performance of our proposed approach, we conducted experiments on high-dimensional datasets with multiple categorical variables. Given that many readily available datasets and data science applications involve such datasets, our experiments demonstrate the effectiveness of our proposed methodology.
In recent years, medical image segmentation has become an important application in the field of computer-aided diagnosis. In this paper, we are the first to propose a new graph convolution-based decoder namely, Cascaded Graph Convolutional Attention Decoder (G-CASCADE), for 2D medical image segmentation. G-CASCADE progressively refines multi-stage feature maps generated by hierarchical transformer encoders with an efficient graph convolution block. The encoder utilizes the self-attention mechanism to capture long-range dependencies, while the decoder refines the feature maps preserving long-range information due to the global receptive fields of the graph convolution block. Rigorous evaluations of our decoder with multiple transformer encoders on five medical image segmentation tasks (i.e., Abdomen organs, Cardiac organs, Polyp lesions, Skin lesions, and Retinal vessels) show that our model outperforms other state-of-the-art (SOTA) methods. We also demonstrate that our decoder achieves better DICE scores than the SOTA CASCADE decoder with 80.8% fewer parameters and 82.3% fewer FLOPs. Our decoder can easily be used with other hierarchical encoders for general-purpose semantic and medical image segmentation tasks.
In this work, we address the problem of real-time dense depth estimation from monocular images for mobile underwater vehicles. We formulate a deep learning model that fuses sparse depth measurements from triangulated features to improve the depth predictions and solve the problem of scale ambiguity. To allow prior inputs of arbitrary sparsity, we apply a dense parameterization method. Our model extends recent state-of-the-art approaches to monocular image based depth estimation, using an efficient encoder-decoder backbone and modern lightweight transformer optimization stage to encode global context. The network is trained in a supervised fashion on the forward-looking underwater dataset, FLSea. Evaluation results on this dataset demonstrate significant improvement in depth prediction accuracy by the fusion of the sparse feature priors. In addition, without any retraining, our method achieves similar depth prediction accuracy on a downward looking dataset we collected with a diver operated camera rig, conducting a survey of a coral reef. The method achieves real-time performance, running at 160 FPS on a laptop GPU and 7 FPS on a single CPU core and is suitable for direct deployment on embedded systems. The implementation of this work is made publicly available at https://github.com/ebnerluca/uw_depth.
TD-MPC is a model-based reinforcement learning (RL) algorithm that performs local trajectory optimization in the latent space of a learned implicit (decoder-free) world model. In this work, we present TD-MPC2: a series of improvements upon the TD-MPC algorithm. We demonstrate that TD-MPC2 improves significantly over baselines across 104 online RL tasks spanning 4 diverse task domains, achieving consistently strong results with a single set of hyperparameters. We further show that agent capabilities increase with model and data size, and successfully train a single 317M parameter agent to perform 80 tasks across multiple task domains, embodiments, and action spaces. We conclude with an account of lessons, opportunities, and risks associated with large TD-MPC2 agents. Explore videos, models, data, code, and more at https://tdmpc2.com
Recently, neural networks have improved MinSum message-passing decoders for low-density parity-check (LDPC) codes by multiplying or adding weights to the messages, where the weights are determined by a neural network. The neural network complexity to determine distinct weights for each edge is high, often limiting the application to relatively short LDPC codes. Furthermore, storing separate weights for every edge and every iteration can be a burden for hardware implementations. To reduce neural network complexity and storage requirements, this paper proposes a family of weight-sharing schemes that use the same weight for edges that have the same check node degree and/or variable node degree. Our simulation results show that node-degree-based weight-sharing can deliver the same performance requiring distinct weights for each node. This paper also combines these degree-specific neural weights with a reconstruction-computation-quantization (RCQ) decoder to produce a weighted RCQ (W-RCQ) decoder. The W-RCQ decoder with node-degree-based weight sharing has a reduced hardware requirement compared with the original RCQ decoder. As an additional contribution, this paper identifies and resolves a gradient explosion issue that can arise when training neural LDPC decoders.
Previous approaches for automatic lay summarisation are exclusively reliant on the source article that, given it is written for a technical audience (e.g., researchers), is unlikely to explicitly define all technical concepts or state all of the background information that is relevant for a lay audience. We address this issue by augmenting eLife, an existing biomedical lay summarisation dataset, with article-specific knowledge graphs, each containing detailed information on relevant biomedical concepts. Using both automatic and human evaluations, we systematically investigate the effectiveness of three different approaches for incorporating knowledge graphs within lay summarisation models, with each method targeting a distinct area of the encoder-decoder model architecture. Our results confirm that integrating graph-based domain knowledge can significantly benefit lay summarisation by substantially increasing the readability of generated text and improving the explanation of technical concepts.
The introduction of DETR represents a new paradigm for object detection. However, its decoder conducts classification and box localization using shared queries and cross-attention layers, leading to suboptimal results. We observe that different regions of interest in the visual feature map are suitable for performing query classification and box localization tasks, even for the same object. Salient regions provide vital information for classification, while the boundaries around them are more favorable for box regression. Unfortunately, such spatial misalignment between these two tasks greatly hinders DETR's training. Therefore, in this work, we focus on decoupling localization and classification tasks in DETR. To achieve this, we introduce a new design scheme called spatially decoupled DETR (SD-DETR), which includes a task-aware query generation module and a disentangled feature learning process. We elaborately design the task-aware query initialization process and divide the cross-attention block in the decoder to allow the task-aware queries to match different visual regions. Meanwhile, we also observe that the prediction misalignment problem for high classification confidence and precise localization exists, so we propose an alignment loss to further guide the spatially decoupled DETR training. Through extensive experiments, we demonstrate that our approach achieves a significant improvement in MSCOCO datasets compared to previous work. For instance, we improve the performance of Conditional DETR by 4.5 AP. By spatially disentangling the two tasks, our method overcomes the misalignment problem and greatly improves the performance of DETR for object detection.
We present DALE, a novel and effective generative Data Augmentation framework for low-resource LEgal NLP. DALE addresses the challenges existing frameworks pose in generating effective data augmentations of legal documents - legal language, with its specialized vocabulary and complex semantics, morphology, and syntax, does not benefit from data augmentations that merely rephrase the source sentence. To address this, DALE, built on an Encoder-Decoder Language Model, is pre-trained on a novel unsupervised text denoising objective based on selective masking - our masking strategy exploits the domain-specific language characteristics of templatized legal documents to mask collocated spans of text. Denoising these spans helps DALE acquire knowledge about legal concepts, principles, and language usage. Consequently, it develops the ability to generate coherent and diverse augmentations with novel contexts. Finally, DALE performs conditional generation to generate synthetic augmentations for low-resource Legal NLP tasks. We demonstrate the effectiveness of DALE on 13 datasets spanning 6 tasks and 4 low-resource settings. DALE outperforms all our baselines, including LLMs, qualitatively and quantitatively, with improvements of 1%-50%.
When translating words referring to the speaker, speech translation (ST) systems should not resort to default masculine generics nor rely on potentially misleading vocal traits. Rather, they should assign gender according to the speakers' preference. The existing solutions to do so, though effective, are hardly feasible in practice as they involve dedicated model re-training on gender-labeled ST data. To overcome these limitations, we propose the first inference-time solution to control speaker-related gender inflections in ST. Our approach partially replaces the (biased) internal language model (LM) implicitly learned by the ST decoder with gender-specific external LMs. Experiments on en->es/fr/it show that our solution outperforms the base models and the best training-time mitigation strategy by up to 31.0 and 1.6 points in gender accuracy, respectively, for feminine forms. The gains are even larger (up to 32.0 and 3.4) in the challenging condition where speakers' vocal traits conflict with their gender.
Natural language serves as the primary mode of communication when an intelligent agent with a physical presence engages with human beings. While a plethora of research focuses on natural language understanding (NLU), encompassing endeavors such as sentiment analysis, intent prediction, question answering, and summarization, the scope of NLU directed at situations necessitating tangible actions by an embodied agent remains limited. The inherent ambiguity and incompleteness inherent in natural language present challenges for intelligent agents striving to decipher human intention. To tackle this predicament head-on, we introduce a novel system known as task and argument grounding for Embodied agents (tagE). At its core, our system employs an inventive neural network model designed to extract a series of tasks from complex task instructions expressed in natural language. Our proposed model adopts an encoder-decoder framework enriched with nested decoding to effectively extract tasks and their corresponding arguments from these intricate instructions. These extracted tasks are then mapped (or grounded) to the robot's established collection of skills, while the arguments find grounding in objects present within the environment. To facilitate the training and evaluation of our system, we have curated a dataset featuring complex instructions. The results of our experiments underscore the prowess of our approach, as it outperforms robust baseline models.
Existing works on Aspect Sentiment Triplet Extraction (ASTE) explicitly focus on developing more efficient fine-tuning techniques for the task. Instead, our motivation is to come up with a generic approach that can improve the downstream performances of multiple ABSA tasks simultaneously. Towards this, we present CONTRASTE, a novel pre-training strategy using CONTRastive learning to enhance the ASTE performance. While we primarily focus on ASTE, we also demonstrate the advantage of our proposed technique on other ABSA tasks such as ACOS, TASD, and AESC. Given a sentence and its associated (aspect, opinion, sentiment) triplets, first, we design aspect-based prompts with corresponding sentiments masked. We then (pre)train an encoder-decoder model by applying contrastive learning on the decoder-generated aspect-aware sentiment representations of the masked terms. For fine-tuning the model weights thus obtained, we then propose a novel multi-task approach where the base encoder-decoder model is combined with two complementary modules, a tagging-based Opinion Term Detector, and a regression-based Triplet Count Estimator. Exhaustive experiments on four benchmark datasets and a detailed ablation study establish the importance of each of our proposed components as we achieve new state-of-the-art ASTE results.
With the recent focus on Large Language Models (LLMs), both StarCoder (Li et al., 2023) and Code Llama (Rozi\`ere et al., 2023) have demonstrated remarkable performance in code generation. However, there is still a need for improvement in code translation functionality with efficient training techniques. In response to this, we introduce SteloCoder, a decoder-only StarCoder-based LLM designed specifically for multi-programming language-to-Python code translation. In particular, SteloCoder achieves C++, C#, JavaScript, Java, or PHP-to-Python code translation without specifying the input programming language. We modified StarCoder model architecture by incorporating a Mixture-of-Experts (MoE) technique featuring five experts and a gating network for multi-task handling. Experts are obtained by StarCoder fine-tuning. Specifically, we use a Low-Rank Adaptive Method (LoRA) technique, limiting each expert size as only 0.06% of number of StarCoder's parameters. At the same time, to enhance training efficiency in terms of time, we adopt curriculum learning strategy and use self-instruct data for efficient fine-tuning. As a result, each expert takes only 6 hours to train on one single 80Gb A100 HBM. With experiments on XLCoST datasets, SteloCoder achieves an average of 73.76 CodeBLEU score in multi-programming language-to-Python translation, surpassing the top performance from the leaderboard by at least 3.5. This accomplishment is attributed to only 45M extra parameters with StarCoder as the backbone and 32 hours of valid training on one 80GB A100 HBM. The source code is release here: https://github.com/sade-adrien/SteloCoder.
In the burgeoning field of natural language processing (NLP), Neural Topic Models (NTMs) , Large Language Models (LLMs) and Diffusion model have emerged as areas of significant research interest. Despite this, NTMs primarily utilize contextual embeddings from LLMs, which are not optimal for clustering or capable for topic based text generation. NTMs have never been combined with diffusion model for text generation. Our study addresses these gaps by introducing a novel framework named Diffusion-Enhanced Topic Modeling using Encoder-Decoder-based LLMs (DeTiME). DeTiME leverages Encoder-Decoder-based LLMs to produce highly clusterable embeddings that could generate topics that exhibit both superior clusterability and enhanced semantic coherence compared to existing methods. Additionally, by exploiting the power of diffusion model, our framework also provides the capability to do topic based text generation. This dual functionality allows users to efficiently produce highly clustered topics and topic based text generation simultaneously. DeTiME's potential extends to generating clustered embeddings as well. Notably, our proposed framework(both encoder-decoder based LLM and diffusion model) proves to be efficient to train and exhibits high adaptability to other LLMs and diffusion model, demonstrating its potential for a wide array of applications.
This work aims efficiently estimating the posterior distribution of kinetic parameters for dynamic positron emission tomography (PET) imaging given a measurement of time of activity curve. Considering the inherent information loss from parametric imaging to measurement space with the forward kinetic model, the inverse mapping is ambiguous. The conventional (but expensive) solution can be the Markov Chain Monte Carlo (MCMC) sampling, which is known to produce unbiased asymptotical estimation. We propose a deep-learning-based framework for efficient posterior estimation. Specifically, we counteract the information loss in the forward process by introducing latent variables. Then, we use a conditional variational autoencoder (CVAE) and optimize its evidence lower bound. The well-trained decoder is able to infer the posterior with a given measurement and the sampled latent variables following a simple multivariate Gaussian distribution. We validate our CVAE-based method using unbiased MCMC as the reference for low-dimensional data (a single brain region) with the simplified reference tissue model.
Automatic aorta segmentation from 3-D medical volumes is an important yet difficult task. Several factors make the problem challenging, e.g. the possibility of aortic dissection or the difficulty with segmenting and annotating the small branches. This work presents a contribution by the MedGIFT team to the SEG.A challenge organized during the MICCAI 2023 conference. We propose a fully automated algorithm based on deep encoder-decoder architecture. The main assumption behind our work is that data preprocessing and augmentation are much more important than the deep architecture, especially in low data regimes. Therefore, the solution is based on a variant of traditional convolutional U-Net. The proposed solution achieved a Dice score above 0.9 for all testing cases with the highest stability among all participants. The method scored 1st, 4th, and 3rd in terms of the clinical evaluation, quantitative results, and volumetric meshing quality, respectively. We freely release the source code, pretrained model, and provide access to the algorithm on the Grand-Challenge platform.