Detecting keywords in texts is important for many text mining applications. Graph-based methods have been commonly used to automatically find the key concepts in texts, however, relevant information provided by embeddings has not been widely used to enrich the graph structure. Here we modeled texts co-occurrence networks, where nodes are words and edges are established either by contextual or semantical similarity. We compared two embedding approaches -- Word2vec and BERT -- to check whether edges created via word embeddings can improve the quality of the keyword extraction method. We found that, in fact, the use of virtual edges can improve the discriminability of co-occurrence networks. The best performance was obtained when we considered low percentages of addition of virtual (embedding) edges. A comparative analysis of structural and dynamical network metrics revealed the degree, PageRank, and accessibility are the metrics displaying the best performance in the model enriched with virtual edges.
3D object detection using LiDAR data is an indispensable component for autonomous driving systems. Yet, only a few LiDAR-based 3D object detection methods leverage segmentation information to further guide the detection process. In this paper, we propose a novel multi-task framework that jointly performs 3D object detection and panoptic segmentation. In our method, the 3D object detection backbone in Bird's-Eye-View (BEV) plane is augmented by the injection of Range-View (RV) feature maps from the 3D panoptic segmentation backbone. This enables the detection backbone to leverage multi-view information to address the shortcomings of each projection view. Furthermore, foreground semantic information is incorporated to ease the detection task by highlighting the locations of each object class in the feature maps. Finally, a new center density heatmap generated based on the instance-level information further guides the detection backbone by suggesting possible box center locations for objects. Our method works with any BEV-based 3D object detection method, and as shown by extensive experiments on the nuScenes dataset, it provides significant performance gains. Notably, the proposed method based on a single-stage CenterPoint 3D object detection network achieved state-of-the-art performance on nuScenes 3D Detection Benchmark with 67.3 NDS.
We present OpenGlue: a free open-source framework for image matching, that uses a Graph Neural Network-based matcher inspired by SuperGlue \cite{sarlin20superglue}. We show that including additional geometrical information, such as local feature scale, orientation, and affine geometry, when available (e.g. for SIFT features), significantly improves the performance of the OpenGlue matcher. We study the influence of the various attention mechanisms on accuracy and speed. We also present a simple architectural improvement by combining local descriptors with context-aware descriptors. The code and pretrained OpenGlue models for the different local features are publicly available.
The problem of effectively exploiting the information multiple data sources has become a relevant but challenging research topic in remote sensing. In this paper, we propose a new approach to exploit the complementarity of two data sources: hyperspectral images (HSIs) and light detection and ranging (LiDAR) data. Specifically, we develop a new dual-channel spatial, spectral and multiscale attention convolutional long short-term memory neural network (called dual-channel A3CLNN) for feature extraction and classification of multisource remote sensing data. Spatial, spectral and multiscale attention mechanisms are first designed for HSI and LiDAR data in order to learn spectral- and spatial-enhanced feature representations, and to represent multiscale information for different classes. In the designed fusion network, a novel composite attention learning mechanism (combined with a three-level fusion strategy) is used to fully integrate the features in these two data sources. Finally, inspired by the idea of transfer learning, a novel stepwise training strategy is designed to yield a final classification result. Our experimental results, conducted on several multisource remote sensing data sets, demonstrate that the newly proposed dual-channel A3CLNN exhibits better feature representation ability (leading to more competitive classification performance) than other state-of-the-art methods.
Contextual knowledge is essential for reducing speech recognition errors on high-valued long-tail words. This paper proposes a novel tree-constrained pointer generator (TCPGen) component that enables end-to-end ASR models to bias towards a list of long-tail words obtained using external contextual information. With only a small overhead in memory use and computation cost, TCPGen can structure thousands of biasing words efficiently into a symbolic prefix-tree and creates a neural shortcut between the tree and the final ASR output to facilitate the recognition of the biasing words. To enhance TCPGen, we further propose a novel minimum biasing word error (MBWE) loss that directly optimises biasing word errors during training, along with a biasing-word-driven language model discounting (BLMD) method during the test. All contextual ASR systems were evaluated on the public Librispeech audiobook corpus and the data from the dialogue state tracking challenges (DSTC) with the biasing lists extracted from the dialogue-system ontology. Consistent word error rate (WER) reductions were achieved with TCPGen, which were particularly significant on the biasing words with around 40\% relative reductions in the recognition error rates. MBWE and BLMD further improved the effectiveness of TCPGen and achieved more significant WER reductions on the biasing words. TCPGen also achieved zero-shot learning of words not in the audio training set with large WER reductions on the out-of-vocabulary words in the biasing list.
We discuss an approach to probabilistic forecasting based on two chained machine-learning steps: a dimensional reduction step that learns a reduction map of predictor information to a low-dimensional space in a manner designed to preserve information about forecast quantities; and a density estimation step that uses the probabilistic machine learning technique of normalizing flows to compute the joint probability density of reduced predictors and forecast quantities. This joint density is then renormalized to produce the conditional forecast distribution. In this method, probabilistic calibration testing plays the role of a regularization procedure, preventing overfitting in the second step, while effective dimensional reduction from the first step is the source of forecast sharpness. We verify the method using a 22-year 1-hour cadence time series of Weather Research and Forecasting (WRF) simulation data of surface wind on a grid.
Speech enhancement (SE) is usually required as a front end to improve the speech quality in noisy environments, while the enhanced speech might not be optimal for automatic speech recognition (ASR) systems due to speech distortion. On the other hand, it was shown that self-supervised pre-training enables the utilization of a large amount of unlabeled noisy data, which is rather beneficial for the noise robustness of ASR. However, the potential of the (optimal) integration of SE and self-supervised pre-training still remains unclear. In order to find an appropriate combination and reduce the impact of speech distortion caused by SE, in this paper we therefore propose a joint pre-training approach for the SE module and the self-supervised model. First, in the pre-training phase the original noisy waveform or the waveform obtained by SE is fed into the self-supervised model to learn the contextual representation, where the quantified clean speech acts as the target. Second, we propose a dual-attention fusion method to fuse the features of noisy and enhanced speeches, which can compensate the information loss caused by separately using individual modules. Due to the flexible exploitation of clean/noisy/enhanced branches, the proposed method turns out to be a generalization of some existing noise-robust ASR models, e.g., enhanced wav2vec2.0. Finally, experimental results on both synthetic and real noisy datasets show that the proposed joint training approach can improve the ASR performance under various noisy settings, leading to a stronger noise robustness.
This work presents a novel approach for speaker diarization to leverage lexical information provided by automatic speech recognition. We propose a speaker diarization system that can incorporate word-level speaker turn probabilities with speaker embeddings into a speaker clustering process to improve the overall diarization accuracy. To integrate lexical and acoustic information in a comprehensive way during clustering, we introduce an adjacency matrix integration for spectral clustering. Since words and word boundary information for word-level speaker turn probability estimation are provided by a speech recognition system, our proposed method works without any human intervention for manual transcriptions. We show that the proposed method improves diarization performance on various evaluation datasets compared to the baseline diarization system using acoustic information only in speaker embeddings.
Most of the existing video face super-resolution (VFSR) methods are trained and evaluated on VoxCeleb1, which is designed specifically for speaker identification and the frames in this dataset are of low quality. As a consequence, the VFSR models trained on this dataset can not output visual-pleasing results. In this paper, we develop an automatic and scalable pipeline to collect a high-quality video face dataset (VFHQ), which contains over $16,000$ high-fidelity clips of diverse interview scenarios. To verify the necessity of VFHQ, we further conduct experiments and demonstrate that VFSR models trained on our VFHQ dataset can generate results with sharper edges and finer textures than those trained on VoxCeleb1. In addition, we show that the temporal information plays a pivotal role in eliminating video consistency issues as well as further improving visual performance. Based on VFHQ, by analyzing the benchmarking study of several state-of-the-art algorithms under bicubic and blind settings. See our project page: https://liangbinxie.github.io/projects/vfhq
Visual understanding requires seamless integration between recognition and reasoning: beyond image-level recognition (e.g., detecting objects), systems must perform concept-level reasoning (e.g., inferring the context of objects and intents of people). However, existing methods only model the image-level features, and do not ground them and reason with background concepts such as knowledge graphs (KGs). In this work, we propose a novel visual question answering method, VQA-GNN, which unifies the image-level information and conceptual knowledge to perform joint reasoning of the scene. Specifically, given a question-image pair, we build a scene graph from the image, retrieve a relevant linguistic subgraph from ConceptNet and visual subgraph from VisualGenome, and unify these three graphs and the question into one joint graph, multimodal semantic graph. Our VQA-GNN then learns to aggregate messages and reason across different modalities captured by the multimodal semantic graph. In the evaluation on the VCR task, our method outperforms the previous scene graph-based Trans-VL models by over 4%, and VQA-GNN-Large, our model that fuses a Trans-VL further improves the state of the art by 2%, attaining the top of the VCR leaderboard at the time of submission. This result suggests the efficacy of our model in performing conceptual reasoning beyond image-level recognition for visual understanding. Finally, we demonstrate that our model is the first work to provide interpretability across visual and textual knowledge domains for the VQA task.