The recently proposed Detection Transformer (DETR) model successfully applies Transformer to objects detection and achieves comparable performance with two-stage object detection frameworks, such as Faster-RCNN. However, DETR suffers from its slow convergence. Training DETR from scratch needs 500 epochs to achieve a high accuracy. To accelerate its convergence, we propose a simple yet effective scheme for improving the DETR framework, namely Spatially Modulated Co-Attention (SMCA) mechanism. The core idea of SMCA is to conduct location-aware co-attention in DETR by constraining co-attention responses to be high near initially estimated bounding box locations. Our proposed SMCA increases DETR's convergence speed by replacing the original co-attention mechanism in the decoder while keeping other operations in DETR unchanged. Furthermore, by integrating multi-head and scale-selection attention designs into SMCA, our fully-fledged SMCA can achieve better performance compared to DETR with a dilated convolution-based backbone (45.6 mAP at 108 epochs vs. 43.3 mAP at 500 epochs). We perform extensive ablation studies on COCO dataset to validate SMCA. Code is released at https://github.com/gaopengcuhk/SMCA-DETR .
The task of video-based commonsense captioning aims to generate event-wise captions and meanwhile provide multiple commonsense descriptions (e.g., attribute, effect and intention) about the underlying event in the video. Prior works explore the commonsense captions by using separate networks for different commonsense types, which is time-consuming and lacks mining the interaction of different commonsense. In this paper, we propose a Hybrid Reasoning Network (HybridNet) to endow the neural networks with the capability of semantic-level reasoning and word-level reasoning. Firstly, we develop multi-commonsense learning for semantic-level reasoning by jointly training different commonsense types in a unified network, which encourages the interaction between the clues of multiple commonsense descriptions, event-wise captions and videos. Then, there are two steps to achieve the word-level reasoning: (1) a memory module records the history predicted sequence from the previous generation processes; (2) a memory-routed multi-head attention (MMHA) module updates the word-level attention maps by incorporating the history information from the memory module into the transformer decoder for word-level reasoning. Moreover, the multimodal features are used to make full use of diverse knowledge for commonsense reasoning. Experiments and abundant analysis on the large-scale Video-to-Commonsense benchmark show that our HybridNet achieves state-of-the-art performance compared with other methods.
Lung nodules are commonly missed in chest radiographs. We propose and evaluate P-AnoGAN, an unsupervised anomaly detection approach for lung nodules in radiographs. P-AnoGAN modifies the fast anomaly detection generative adversarial network (f-AnoGAN) by utilizing a progressive GAN and a convolutional encoder-decoder-encoder pipeline. Model training uses only unlabelled healthy lung patches extracted from the Indiana University Chest X-Ray Collection. External validation and testing are performed using healthy and unhealthy patches extracted from the ChestX-ray14 and Japanese Society for Radiological Technology datasets, respectively. Our model robustly identifies patches containing lung nodules in external validation and test data with ROC-AUC of 91.17% and 87.89%, respectively. These results show unsupervised methods may be useful in challenging tasks such as lung nodule detection in radiographs.
For an image with multiple scene texts, different people may be interested in different text information. Current text-aware image captioning models are not able to generate distinctive captions according to various information needs. To explore how to generate personalized text-aware captions, we define a new challenging task, namely Question-controlled Text-aware Image Captioning (Qc-TextCap). With questions as control signals, this task requires models to understand questions, find related scene texts and describe them together with objects fluently in human language. Based on two existing text-aware captioning datasets, we automatically construct two datasets, ControlTextCaps and ControlVizWiz to support the task. We propose a novel Geometry and Question Aware Model (GQAM). GQAM first applies a Geometry-informed Visual Encoder to fuse region-level object features and region-level scene text features with considering spatial relationships. Then, we design a Question-guided Encoder to select the most relevant visual features for each question. Finally, GQAM generates a personalized text-aware caption with a Multimodal Decoder. Our model achieves better captioning performance and question answering ability than carefully designed baselines on both two datasets. With questions as control signals, our model generates more informative and diverse captions than the state-of-the-art text-aware captioning model. Our code and datasets are publicly available at https://github.com/HAWLYQ/Qc-TextCap.
Previous work has shown that neural architectures are able to perform optical music recognition (OMR) on monophonic and homophonic music with high accuracy. However, piano and orchestral scores frequently exhibit polyphonic passages, which add a second dimension to the task. Monophonic and homophonic music can be described as homorhythmic, or having a single musical rhythm. Polyphonic music, on the other hand, can be seen as having multiple rhythmic sequences, or voices, concurrently. We first introduce a workflow for creating large-scale polyphonic datasets suitable for end-to-end recognition from sheet music publicly available on the MuseScore forum. We then propose two novel formulations for end-to-end polyphonic OMR -- one treating the problem as a type of multi-task binary classification, and the other treating it as multi-sequence detection. Building upon the encoder-decoder architecture and an image encoder proposed in past work on end-to-end OMR, we propose two novel decoder models -- FlagDecoder and RNNDecoder -- that correspond to the two formulations. Finally, we compare the empirical performance of these end-to-end approaches to polyphonic OMR and observe a new state-of-the-art performance with our multi-sequence detection decoder, RNNDecoder.
Improving controllability or the ability to manipulate one or more attributes of the generated data has become a topic of interest in the context of deep generative models of music. Recent attempts in this direction have relied on learning disentangled representations from data such that the underlying factors of variation are well separated. In this paper, we focus on the relationship between disentanglement and controllability by conducting a systematic study using different supervised disentanglement learning algorithms based on the Variational Auto-Encoder (VAE) architecture. Our experiments show that a high degree of disentanglement can be achieved by using different forms of supervision to train a strong discriminative encoder. However, in the absence of a strong generative decoder, disentanglement does not necessarily imply controllability. The structure of the latent space with respect to the VAE-decoder plays an important role in boosting the ability of a generative model to manipulate different attributes. To this end, we also propose methods and metrics to help evaluate the quality of a latent space with respect to the afforded degree of controllability.
Soft spelling errors are a class of spelling mistakes that is widespread among native Arabic speakers and foreign learners alike. Some of these errors are typographical in nature. They occur due to orthographic variations of some Arabic letters and the complex rules that dictate their correct usage. Many people forgo these rules, and given the identical phonetic sounds, they often confuse such letters. In this paper, we propose a bidirectional long short-term memory network that corrects this class of errors. We develop, train, evaluate, and compare a set of BiLSTM networks. We approach the spelling correction problem at the character level. We handle Arabic texts from both classical and modern standard Arabic. We treat the problem as a one-to-one sequence transcription problem. Since the soft Arabic errors class encompasses omission and addition mistakes, to preserve the one-to-one sequence transcription, we propose a simple low-resource yet effective technique that maintains the one-to-one sequencing and avoids using a costly encoder-decoder architecture. We train the BiLSTM models to correct the spelling mistakes using transformed input and stochastic error injection approaches. We recommend a configuration that has two BiLSTM layers, uses the dropout regularization, and is trained using the latter training approach with error injection rate of 40%. The best model corrects 96.4% of the injected errors and achieves a low character error rate of 1.28% on a real test set of soft spelling mistakes.
3D skeleton-based motion prediction and activity recognition are two interwoven tasks in human behaviour analysis. In this work, we propose a motion context modeling methodology that provides a new way to combine the advantages of both graph convolutional neural networks and recurrent neural networks for joint human motion prediction and activity recognition. Our approach is based on using an LSTM encoder-decoder and a non-local feature extraction attention mechanism to model the spatial correlation of human skeleton data and temporal correlation among motion frames. The proposed network can easily include two output branches, one for Activity Recognition and one for Future Motion Prediction, which can be jointly trained for enhanced performance. Experimental results on Human 3.6M, CMU Mocap and NTU RGB-D datasets show that our proposed approach provides the best prediction capability among baseline LSTM-based methods, while achieving comparable performance to other state-of-the-art methods.
The problem of end-to-end learning of a communication system using an autoencoder -- consisting of an encoder, channel, and decoder modeled using neural networks -- has recently been shown to be an effective approach. A challenge faced in the practical adoption of this learning approach is that under changing channel conditions (e.g. a wireless link), it requires frequent retraining of the autoencoder in order to maintain a low decoding error rate. Since retraining is both time consuming and requires a large number of samples, it becomes impractical when the channel distribution is changing quickly. We propose to address this problem using a fast and sample-efficient (few-shot) domain adaptation method that does not change the encoder and decoder networks. Different from conventional training-time unsupervised or semi-supervised domain adaptation, here we have a trained autoencoder from a source distribution that we want to adapt (at test time) to a target distribution using only a small labeled dataset, and no unlabeled data. We focus on a generative channel model based on the Gaussian mixture density network (MDN), and propose a regularized, parameter-efficient adaptation of the MDN using a set of affine transformations. The learned affine transformations are then used to design an optimal transformation at the decoder input to compensate for the distribution shift, and effectively present to the decoder inputs close to the source distribution. Experiments on many simulated distribution changes common to the wireless setting, and a real mmWave FPGA testbed demonstrate the effectiveness of our method at adaptation using very few target domain samples. The code for our work can be found at: https://github.com/jayaram-r/domain-adaptation-autoencoder.
As a fundamental building block in computer vision, edges can be categorised into four types according to the discontinuity in surface-Reflectance, Illumination, surface-Normal or Depth. While great progress has been made in detecting generic or individual types of edges, it remains under-explored to comprehensively study all four edge types together. In this paper, we propose a novel neural network solution, RINDNet, to jointly detect all four types of edges. Taking into consideration the distinct attributes of each type of edges and the relationship between them, RINDNet learns effective representations for each of them and works in three stages. In stage I, RINDNet uses a common backbone to extract features shared by all edges. Then in stage II it branches to prepare discriminative features for each edge type by the corresponding decoder. In stage III, an independent decision head for each type aggregates the features from previous stages to predict the initial results. Additionally, an attention module learns attention maps for all types to capture the underlying relations between them, and these maps are combined with initial results to generate the final edge detection results. For training and evaluation, we construct the first public benchmark, BSDS-RIND, with all four types of edges carefully annotated. In our experiments, RINDNet yields promising results in comparison with state-of-the-art methods. Additional analysis is presented in supplementary material.
In medical imaging, there are clinically relevant segmentation tasks where the output mask is a projection to a subset of input image dimensions. In this work, we propose a novel convolutional neural network architecture that can effectively learn to produce a lower-dimensional segmentation mask than the input image. The network restores encoded representation only in a subset of input spatial dimensions and keeps the representation unchanged in the others. The newly proposed projective skip-connections allow linking the encoder and decoder in a UNet-like structure. We evaluated the proposed method on two clinically relevant tasks in retinal Optical Coherence Tomography (OCT): geographic atrophy and retinal blood vessel segmentation. The proposed method outperformed the current state-of-the-art approaches on all the OCT datasets used, consisting of 3D volumes and corresponding 2D en-face masks. The proposed architecture fills the methodological gap between image classification and ND image segmentation.
Text style transfer aims to alter the style (e.g., sentiment) of a sentence while preserving its content. A common approach is to map a given sentence to content representation that is free of style, and the content representation is fed to a decoder with a target style. Previous methods in filtering style completely remove tokens with style at the token level, which incurs the loss of content information. In this paper, we propose to enhance content preservation by implicitly removing the style information of each token with reverse attention, and thereby retain the content. Furthermore, we fuse content information when building the target style representation, making it dynamic with respect to the content. Our method creates not only style-independent content representation, but also content-dependent style representation in transferring style. Empirical results show that our method outperforms the state-of-the-art baselines by a large margin in terms of content preservation. In addition, it is also competitive in terms of style transfer accuracy and fluency.
Current one-stage methods for visual grounding encode the language query as one holistic sentence embedding before fusion with visual feature. Such a formulation does not treat each word of a query sentence on par when modeling language to visual attention, therefore prone to neglect words which are less important for sentence embedding but critical for visual grounding. In this paper we propose Word2Pix: a one-stage visual grounding network based on encoder-decoder transformer architecture that enables learning for textual to visual feature correspondence via word to pixel attention. The embedding of each word from the query sentence is treated alike by attending to visual pixels individually instead of single holistic sentence embedding. In this way, each word is given equivalent opportunity to adjust the language to vision attention towards the referent target through multiple stacks of transformer decoder layers. We conduct the experiments on RefCOCO, RefCOCO+ and RefCOCOg datasets and the proposed Word2Pix outperforms existing one-stage methods by a notable margin. The results obtained also show that Word2Pix surpasses two-stage visual grounding models, while at the same time keeping the merits of one-stage paradigm namely end-to-end training and real-time inference speed intact.
There is unison is the scientific community about human induced climate change. Despite this, we see the web awash with claims around climate change scepticism, thus driving the need for fact checking them but at the same time providing an explanation and justification for the fact check. Scientists and experts have been trying to address it by providing manually written feedback for these claims. In this paper, we try to aid them by automating generating explanation for a predicted veracity label for a claim by deploying the approach used in open domain question answering of a fusion in decoder augmented with retrieved supporting passages from an external knowledge. We experiment with different knowledge sources, retrievers, retriever depths and demonstrate that even a small number of high quality manually written explanations can help us in generating good explanations.
Originating in the Renaissance and burgeoning in the digital era, tablatures are a commonly used music notation system which provides explicit representations of instrument fingerings rather than pitches. GuitarPro has established itself as a widely used tablature format and software enabling musicians to edit and share songs for musical practice, learning, and composition. In this work, we present DadaGP, a new symbolic music dataset comprising 26,181 song scores in the GuitarPro format covering 739 musical genres, along with an accompanying tokenized format well-suited for generative sequence models such as the Transformer. The tokenized format is inspired by event-based MIDI encodings, often used in symbolic music generation models. The dataset is released with an encoder/decoder which converts GuitarPro files to tokens and back. We present results of a use case in which DadaGP is used to train a Transformer-based model to generate new songs in GuitarPro format. We discuss other relevant use cases for the dataset (guitar-bass transcription, music style transfer and artist/genre classification) as well as ethical implications. DadaGP opens up the possibility to train GuitarPro score generators, fine-tune models on custom data, create new styles of music, AI-powered songwriting apps, and human-AI improvisation.
Although attention-based Neural Machine Translation has achieved remarkable progress in recent layers, it still suffers from issue of making insufficient use of the output of each layer. In transformer, it only uses the top layer of encoder and decoder in the subsequent process, which makes it impossible to take advantage of the useful information in other layers. To address this issue, we propose a residual tree aggregation of layers for Transformer(RTAL), which helps to fuse information across layers. Specifically, we try to fuse the information across layers by constructing a post-order binary tree. In additional to the last node, we add the residual connection to the process of generating child nodes. Our model is based on the Neural Machine Translation model Transformer and we conduct our experiments on WMT14 English-to-German and WMT17 English-to-France translation tasks. Experimental results across language pairs show that the proposed approach outperforms the strong baseline model significantly
Semantic segmentation requires per-pixel prediction for a given image. Typically, the output resolution of a segmentation network is severely reduced due to the downsampling operations in the CNN backbone. Most previous methods employ upsampling decoders to recover the spatial resolution. Various decoders were designed in the literature. Here, we propose a novel decoder, termed dynamic neural representational decoder (NRD), which is simple yet significantly more efficient. As each location on the encoder's output corresponds to a local patch of the semantic labels, in this work, we represent these local patches of labels with compact neural networks. This neural representation enables our decoder to leverage the smoothness prior in the semantic label space, and thus makes our decoder more efficient. Furthermore, these neural representations are dynamically generated and conditioned on the outputs of the encoder networks. The desired semantic labels can be efficiently decoded from the neural representations, resulting in high-resolution semantic segmentation predictions. We empirically show that our proposed decoder can outperform the decoder in DeeplabV3+ with only 30% computational complexity, and achieve competitive performance with the methods using dilated encoders with only 15% computation. Experiments on the Cityscapes, ADE20K, and PASCAL Context datasets demonstrate the effectiveness and efficiency of our proposed method.
Short-term forecasting of passenger flow is critical for transit management and crowd regulation. Spatial dependencies, temporal dependencies, inter-station correlations driven by other latent factors, and exogenous factors bring challenges to the short-term forecasts of passenger flow of urban rail transit networks. An innovative deep learning approach, Multi-Graph Convolutional-Recurrent Neural Network (MGC-RNN) is proposed to forecast passenger flow in urban rail transit systems to incorporate these complex factors. We propose to use multiple graphs to encode the spatial and other heterogenous inter-station correlations. The temporal dynamics of the inter-station correlations are also modeled via the proposed multi-graph convolutional-recurrent neural network structure. Inflow and outflow of all stations can be collectively predicted with multiple time steps ahead via a sequence to sequence(seq2seq) architecture. The proposed method is applied to the short-term forecasts of passenger flow in Shenzhen Metro, China. The experimental results show that MGC-RNN outperforms the benchmark algorithms in terms of forecasting accuracy. Besides, it is found that the inter-station driven by network distance, network structure, and recent flow patterns are significant factors for passenger flow forecasting. Moreover, the architecture of LSTM-encoder-decoder can capture the temporal dependencies well. In general, the proposed framework could provide multiple views of passenger flow dynamics for fine prediction and exhibit a possibility for multi-source heterogeneous data fusion in the spatiotemporal forecast tasks.
We present a framework for the unsupervised learning of neurosymbolic encoders, which are encoders obtained by composing neural networks with symbolic programs from a domain-specific language. Our framework naturally incorporates symbolic expert knowledge into the learning process, which leads to more interpretable and factorized latent representations compared to fully neural encoders. We integrate modern program synthesis techniques with the variational autoencoding (VAE) framework, in order to learn a neurosymbolic encoder in conjunction with a standard decoder. The programmatic descriptions from our encoders can benefit many analysis workflows, such as in behavior modeling where interpreting agent actions and movements is important. We evaluate our method on learning latent representations for real-world trajectory data from animal biology and sports analytics. We show that our approach offers significantly better separation of meaningful categories than standard VAEs and leads to practical gains on downstream analysis tasks, such as for behavior classification.
Image captioning is a task in the field of Artificial Intelligence that merges between computer vision and natural language processing. It is responsible for generating legends that describe images, and has various applications like descriptions used by assistive technology or indexing images (for search engines for instance). This makes it a crucial topic in AI that is undergoing a lot of research. This task however, like many others, is trained on large images labeled via human annotation, which can be very cumbersome: it needs manual effort, both financial and temporal costs, it is error-prone and potentially difficult to execute in some cases (e.g. medical images). To mitigate the need for labels, we attempt to use self-supervised learning, a type of learning where models use the data contained within the images themselves as labels. It is challenging to accomplish though, since the task is two-fold: the images and captions come from two different modalities and usually handled by different types of networks. It is thus not obvious what a completely self-supervised solution would look like. How it would achieve captioning in a comparable way to how self-supervision is applied today on image recognition tasks is still an ongoing research topic. In this project, we are using an encoder-decoder architecture where the encoder is a convolutional neural network (CNN) trained on OpenImages dataset and learns image features in a self-supervised fashion using the rotation pretext task. The decoder is a Long Short-Term Memory (LSTM), and it is trained, along within the image captioning model, on MS COCO dataset and is responsible of generating captions. Our GitHub repository can be found: https://github.com/elhagry1/SSL_ImageCaptioning_RotationPrediction