Recently, the dominant DETR-based approaches apply central-concept spatial prior to accelerate Transformer detector convergency. These methods gradually refine the reference points to the center of target objects and imbue object queries with the updated central reference information for spatially conditional attention. However, centralizing reference points may severely deteriorate queries' saliency and confuse detectors due to the indiscriminative spatial prior. To bridge the gap between the reference points of salient queries and Transformer detectors, we propose SAlient Point-based DETR (SAP-DETR) by treating object detection as a transformation from salient points to instance objects. In SAP-DETR, we explicitly initialize a query-specific reference point for each object query, gradually aggregate them into an instance object, and then predict the distance from each side of the bounding box to these points. By rapidly attending to query-specific reference region and other conditional extreme regions from the image features, SAP-DETR can effectively bridge the gap between the salient point and the query-based Transformer detector with a significant convergency speed. Our extensive experiments have demonstrated that SAP-DETR achieves 1.4 times convergency speed with competitive performance. Under the standard training scheme, SAP-DETR stably promotes the SOTA approaches by 1.0 AP. Based on ResNet-DC-101, SAP-DETR achieves 46.9 AP.
Dialogue understanding tasks often necessitate abundant annotated data to achieve good performance and that presents challenges in low-resource settings. To alleviate this barrier, we explore few-shot data augmentation for dialogue understanding by prompting large pre-trained language models and present a novel approach that iterates on augmentation quality by applying weakly-supervised filters. We evaluate our methods on the emotion and act classification tasks in DailyDialog and the intent classification task in Facebook Multilingual Task-Oriented Dialogue. Models fine-tuned on our augmented data mixed with few-shot ground truth data are able to approach or surpass existing state-of-the-art performance on both datasets. For DailyDialog specifically, using 10% of the ground truth data we outperform the current state-of-the-art model which uses 100% of the data.
Video anomaly detection (VAD) remains a challenging task in the pattern recognition community due to the ambiguity and diversity of abnormal events. Existing deep learning-based VAD methods usually leverage proxy tasks to learn the normal patterns and discriminate the instances that deviate from such patterns as abnormal. However, most of them do not take full advantage of spatial-temporal correlations among video frames, which is critical for understanding normal patterns. In this paper, we address unsupervised VAD by learning the evolution regularity of appearance and motion in the long and short-term and exploit the spatial-temporal correlations among consecutive frames in normal videos more adequately. Specifically, we proposed to utilize the spatiotemporal long short-term memory (ST-LSTM) to extract and memorize spatial appearances and temporal variations in a unified memory cell. In addition, inspired by the generative adversarial network, we introduce a discriminator to perform adversarial learning with the ST-LSTM to enhance the learning capability. Experimental results on standard benchmarks demonstrate the effectiveness of spatial-temporal correlations for unsupervised VAD. Our method achieves competitive performance compared to the state-of-the-art methods with AUCs of 96.7%, 87.8%, and 73.1% on the UCSD Ped2, CUHK Avenue, and ShanghaiTech, respectively.
There has been an increasing interest in integrating physics knowledge and machine learning for modeling dynamical systems. However, very limited studies have been conducted on seismic wave modeling tasks. A critical challenge is that these geophysical problems are typically defined in large domains (i.e., semi-infinite), which leads to high computational cost. In this paper, we present a novel physics-informed neural network (PINN) model for seismic wave modeling in semi-infinite domain without the nedd of labeled data. In specific, the absorbing boundary condition is introduced into the network as a soft regularizer for handling truncated boundaries. In terms of computational efficiency, we consider a sequential training strategy via temporal domain decomposition to improve the scalability of the network and solution accuracy. Moreover, we design a novel surrogate modeling strategy for parametric loading, which estimates the wave propagation in semin-infinite domain given the seismic loading at different locations. Various numerical experiments have been implemented to evaluate the performance of the proposed PINN model in the context of forward modeling of seismic wave propagation. In particular, we define diverse material distributions to test the versatility of this approach. The results demonstrate excellent solution accuracy under distinctive scenarios.
End-to-End deep learning has shown promising results for speech enhancement tasks, such as noise suppression, dereverberation, and speech separation. However, most state-of-the-art methods for echo cancellation are either classical DSP-based or hybrid DSP-ML algorithms. Components such as the delay estimator and adaptive linear filter are based on traditional signal processing concepts, and deep learning algorithms typically only serve to replace the non-linear residual echo suppressor. This paper introduces an end-to-end echo cancellation network with a streaming cross-attention alignment (SCA). Our proposed method can handle unaligned inputs without requiring external alignment and generate high-quality speech without echoes. At the same time, the end-to-end algorithm simplifies the current echo cancellation pipeline for time-variant echo path cases. We test our proposed method on the ICASSP2022 and Interspeech2021 Microsoft deep echo cancellation challenge evaluation dataset, where our method outperforms some of the other hybrid and end-to-end methods.
Reconfigurable intelligent surface (RIS) is a revolutionary technology for sixth-generation (6G) networks owing to its ability to manipulate wireless environments. As a frequency-selective device, RIS can only effectively shape the propagation of signals within a certain frequency band. Due to this frequency-selective property, the deployment of RIS in cellular networks will introduce a complicated base station (BS)-RIS-user association issue since adjacent BSs operate at different frequency bands. In this paper, with the consideration of the frequency-selective characteristics of RIS, we aim to jointly optimize BS-RIS-user association, active beamforming at BSs, and passive beamforming of RIS to maximize the sum-rate of a RIS-assisted cellular network. We first leverage $l_0$-norm to efficiently integrate BS-RIS-user association with active and passive beamforming. Then, we adopt fractional programming (FP) and block coordinate descent (BCD) methods to deal with logarithmic and fractional parts and decouple the joint association and beamforming design problem into several sub-problems. Efficient algorithms which combine $l_0$-norm approximation, majorization-minimization (MM), and alternating direction method of multipliers (ADMM) are developed to alternately solve the sub-problems. Extensive simulation results illustrate the importance of BS-RIS-user association optimization in RIS-assisted cellular networks and verify the effectiveness of the proposed joint association and beamforming design algorithm.
Biological functions of RNAs are determined by their three-dimensional (3D) structures. Thus, given the limited number of experimentally determined RNA structures, the prediction of RNA structures will facilitate elucidating RNA functions and RNA-targeted drug discovery, but remains a challenging task. In this work, we propose a Graph Neural Network (GNN)-based scoring function trained only with the atomic types and coordinates on limited solved RNA 3D structures for distinguishing accurate structural models. The proposed Physics-aware Multiplex Graph Neural Network (PaxNet) separately models the local and non-local interactions inspired by molecular mechanics. Furthermore, PaxNet contains an attention-based fusion module that learns the individual contribution of each interaction type for the final prediction. We rigorously evaluate the performance of PaxNet on two benchmarks and compare it with several state-of-the-art baselines. The results show that PaxNet significantly outperforms all the baselines overall, and demonstrate the potential of PaxNet for improving the 3D structure modeling of RNA and other macromolecules.
Prefix-tuning, or more generally continuous prompt tuning, has become an essential paradigm of parameter-efficient transfer learning. Using a large pre-trained language model (PLM), prefix-tuning can obtain strong performance by training only a small portion of parameters. In this paper, we propose to understand and further develop prefix-tuning through the kernel lens. Specifically, we make an analogy between \textit{prefixes} and \textit{inducing variables} in kernel methods and hypothesize that \textit{prefixes} serving as \textit{inducing variables} would improve their overall mechanism. From the kernel estimator perspective, we suggest a new variant of prefix-tuning -- \textit{inducer-tuning}, which shares the exact mechanism as prefix-tuning while leveraging the residual form found in adapter-tuning. This mitigates the initialization issue in prefix-tuning. Through comprehensive empirical experiments on natural language understanding and generation tasks, we demonstrate that inducer-tuning can close the performance gap between prefix-tuning and fine-tuning.
Previous work on pun generation commonly begins with a given pun word (a pair of homophones for heterographic pun generation and a polyseme for homographic pun generation) and seeks to generate an appropriate pun. While this may enable efficient pun generation, we believe that a pun is most entertaining if it fits appropriately within a given context, e.g., a given situation or dialogue. In this work, we propose a new task, context-situated pun generation, where a specific context represented by a set of keywords is provided, and the task is to first identify suitable pun words that are appropriate for the context, then generate puns based on the context keywords and the identified pun words. We collect CUP (Context-sitUated Pun), containing 4.5k tuples of context words and pun pairs. Based on the new data and setup, we propose a pipeline system for context-situated pun generation, including a pun word retrieval module that identifies suitable pun words for a given context, and a generation module that generates puns from context keywords and pun words. Human evaluation shows that 69% of our top retrieved pun words can be used to generate context-situated puns, and our generation module yields successful puns 31% of the time given a plausible tuple of context words and pun pair, almost tripling the yield of a state-of-the-art pun generation model. With an end-to-end evaluation, our pipeline system with the top-1 retrieved pun pair for a given context can generate successful puns 40% of the time, better than all other modeling variations but 32% lower than the human success rate. This highlights the difficulty of the task, and encourages more research in this direction.
The tasks of humor understanding and generation are challenging and subjective even for humans, requiring commonsense and real-world knowledge to master. Puns, in particular, add the challenge of fusing that knowledge with the ability to interpret lexical-semantic ambiguity. In this paper, we present the ExPUNations (ExPUN) dataset, in which we augment an existing dataset of puns with detailed crowdsourced annotations of keywords denoting the most distinctive words that make the text funny, pun explanations describing why the text is funny, and fine-grained funniness ratings. This is the first humor dataset with such extensive and fine-grained annotations specifically for puns. Based on these annotations, we propose two tasks: explanation generation to aid with pun classification and keyword-conditioned pun generation, to challenge the current state-of-the-art natural language understanding and generation models' ability to understand and generate humor. We showcase that the annotated keywords we collect are helpful for generating better novel humorous texts in human evaluation, and that our natural language explanations can be leveraged to improve both the accuracy and robustness of humor classifiers.