In this paper, we try to uncover the second-order essence of several first-order optimization methods. For Nesterov Accelerated Gradient, we rigorously prove that the algorithm makes use of the difference between past and current gradients, thus approximates the Hessian and accelerates the training. For adaptive methods, we related Adam and Adagrad to a powerful technique in computation statistics---Natural Gradient Descent. These adaptive methods can in fact be treated as relaxations of NGD with only a slight difference lying in the square root of the denominator in the update rules. Skeptical about the effect of such difference, we design a new algorithm---AdaSqrt, which removes the square root in the denominator and scales the learning rate by sqrt(T). Surprisingly, our new algorithm is comparable to various first-order methods(such as SGD and Adam) on MNIST and even beats Adam on CIFAR-10! This phenomenon casts doubt on the convention view that the square root is crucial and training without it will lead to terrible performance. As far as we have concerned, so long as the algorithm tries to explore second or even higher information of the loss surface, then proper scaling of the learning rate alone will guarantee fast training and good generalization performance. To the best of our knowledge, this is the first paper that seriously considers the necessity of square root among all adaptive methods. We believe that our work can shed light on the importance of higher-order information and inspire the design of more powerful algorithms in the future.
Fact-checking is an essential tool to mitigate the spread of misinformation and disinformation, however, it has been often explored to verify formal single-sentence claims instead of casual conversational claims. To study the problem, we introduce the task of fact-checking in dialogue. We construct DialFact, a testing benchmark dataset of 22,245 annotated conversational claims, paired with pieces of evidence from Wikipedia. There are three sub-tasks in DialFact: 1) Verifiable claim detection task distinguishes whether a response carries verifiable factual information; 2) Evidence retrieval task retrieves the most relevant Wikipedia snippets as evidence; 3) Claim verification task predicts a dialogue response to be supported, refuted, or not enough information. We found that existing fact-checking models trained on non-dialogue data like FEVER fail to perform well on our task, and thus, we propose a simple yet data-efficient solution to effectively improve fact-checking performance in dialogue. We point out unique challenges in DialFact such as handling the colloquialisms, coreferences, and retrieval ambiguities in the error analysis to shed light on future research in this direction.
The 3D LiDAR place recognition aims to estimate a coarse localization in a previously seen environment based on a single scan from a rotating 3D LiDAR sensor. The existing solutions to this problem include hand-crafted point cloud descriptors (e.g., ScanContext, M2DP, LiDAR IRIS) and deep learning-based solutions (e.g., PointNetVLAD, PCAN, LPDNet, DAGC, MinkLoc3D), which are often only evaluated on accumulated 2D scans from the Oxford RobotCar dataset. We introduce MinkLoc3D-SI, a sparse convolution-based solution that utilizes spherical coordinates of 3D points and processes the intensity of 3D LiDAR measurements, improving the performance when a single 3D LiDAR scan is used. Our method integrates the improvements typical for hand-crafted descriptors (like ScanContext) with the most efficient 3D sparse convolutions (MinkLoc3D). Our experiments show improved results on single scans from 3D LiDARs (USyd Campus dataset) and great generalization ability (KITTI dataset). Using intensity information on accumulated 2D scans (RobotCar Intensity dataset) improves the performance, even though spherical representation doesn't produce a noticeable improvement. As a result, MinkLoc3D-SI is suited for single scans obtained from a 3D LiDAR, making it applicable in autonomous vehicles.
While self-supervised representation learning (SSL) has proved to be effective in the large model, there is still a huge gap between the SSL and supervised method in the lightweight model when following the same solution. We delve into this problem and find that the lightweight model is prone to collapse in semantic space when simply performing instance-wise contrast. To address this issue, we propose a relation-wise contrastive paradigm with Relation Knowledge Distillation (ReKD). We introduce a heterogeneous teacher to explicitly mine the semantic information and transferring a novel relation knowledge to the student (lightweight model). The theoretical analysis supports our main concern about instance-wise contrast and verify the effectiveness of our relation-wise contrastive learning. Extensive experimental results also demonstrate that our method achieves significant improvements on multiple lightweight models. Particularly, the linear evaluation on AlexNet obviously improves the current state-of-art from 44.7% to 50.1%, which is the first work to get close to the supervised 50.5%. Code will be made available.
Throughout the Covid-19 pandemic, a significant amount of effort had been put into developing techniques that predict the number of infections under various assumptions about the public policy and non-pharmaceutical interventions. While both the available data and the sophistication of the AI models and available computing power exceed what was available in previous years, the overall success of prediction approaches was very limited. In this paper, we start from prediction algorithms proposed for XPrize Pandemic Response Challenge and consider several directions that might allow their improvement. Then, we investigate their performance over medium-term predictions extending over several months. We find that augmenting the algorithms with additional information about the culture of the modeled region, incorporating traditional compartmental models and up-to-date deep learning architectures can improve the performance for short term predictions, the accuracy of medium-term predictions is still very low and a significant amount of future research is needed to make such models a reliable component of a public policy toolbox.
The multi-modal salient object detection model based on RGB-D information has better robustness in the real world. However, it remains nontrivial to better adaptively balance effective multi-modal information in the feature fusion phase. In this letter, we propose a novel gated recoding network (GRNet) to evaluate the information validity of the two modes, and balance their influence. Our framework is divided into three phases: perception phase, recoding mixing phase and feature integration phase. First, A perception encoder is adopted to extract multi-level single-modal features, which lays the foundation for multi-modal semantic comparative analysis. Then, a modal-adaptive gate unit (MGU) is proposed to suppress the invalid information and transfer the effective modal features to the recoding mixer and the hybrid branch decoder. The recoding mixer is responsible for recoding and mixing the balanced multi-modal information. Finally, the hybrid branch decoder completes the multi-level feature integration under the guidance of an optional edge guidance stream (OEGS). Experiments and analysis on eight popular benchmarks verify that our framework performs favorably against 9 state-of-art methods.
In recent years, channel attention mechanism is widely investigated for its great potential in improving the performance of deep convolutional neural networks (CNNs). However, in most existing methods, only the output of the adjacent convolution layer is fed to the attention layer for calculating the channel weights. Information from other convolution layers is ignored. With these observations, a simple strategy, named Bridge Attention Net (BA-Net), is proposed for better channel attention mechanisms. The main idea of this design is to bridge the outputs of the previous convolution layers through skip connections for channel weights generation. BA-Net can not only provide richer features to calculate channel weight when feedforward, but also multiply paths of parameters updating when backforward. Comprehensive evaluation demonstrates that the proposed approach achieves state-of-the-art performance compared with the existing methods in regards to accuracy and speed. Bridge Attention provides a fresh perspective on the design of neural network architectures and shows great potential in improving the performance of the existing channel attention mechanisms. The code is available at \url{https://github.com/zhaoy376/Attention-mechanism
Mixed communities of organisms are found in many environments (from the human gut to marine ecosystems) and can have profound impact on human health and the environment. Metagenomics studies the genomic material of such communities through high-throughput sequencing that yields DNA subsequences for subsequent analysis. A fundamental problem in the standard workflow, called binning, is to discover clusters, of genomic subsequences, associated with the unknown constituent organisms. Inherent noise in the subsequences, various biological constraints that need to be imposed on them and the skewed cluster size distribution exacerbate the difficulty of this unsupervised learning problem. In this paper, we present a new formulation using a graph where the nodes are subsequences and edges represent homophily information. In addition, we model biological constraints providing heterophilous signal about nodes that cannot be clustered together. We solve the binning problem by developing new algorithms for (i) graph representation learning that preserves both homophily relations and heterophily constraints (ii) constraint-based graph clustering method that addresses the problems of skewed cluster size distribution. Extensive experiments, on real and synthetic datasets, demonstrate that our approach, called RepBin, outperforms a wide variety of competing methods. Our constraint-based graph representation learning and clustering methods, that may be useful in other domains as well, advance the state-of-the-art in both metagenomics binning and graph representation learning.
This paper presents a new task of predicting the coverage of a text document for relation extraction (RE): does the document contain many relational tuples for a given entity? Coverage predictions are useful in selecting the best documents for knowledge base construction with large input corpora. To study this problem, we present a dataset of 31,366 diverse documents for 520 entities. We analyze the correlation of document coverage with features like length, entity mention frequency, Alexa rank, language complexity and information retrieval scores. Each of these features has only moderate predictive power. We employ methods combining features with statistical models like TF-IDF and language models like BERT. The model combining features and BERT, HERB, achieves an F1 score of up to 46%. We demonstrate the utility of coverage predictions on two use cases: KB construction and claim refutation.
While we pay attention to the latest advances in clinical natural language processing (NLP), we can notice some resistance in the clinical and translational research community to adopt NLP models due to limited transparency, Interpretability and usability. Built upon our previous work, in this study, we proposed an open natural language processing development framework and evaluated it through the implementation of NLP algorithms for the National COVID Cohort Collaborative (N3C). Based on the interests in information extraction from COVID-19 related clinical notes, our work includes 1) an open data annotation process using COVID-19 signs and symptoms as the use case, 2) a community-driven ruleset composing platform, and 3) a synthetic text data generation workflow to generate texts for information extraction tasks without involving human subjects. The generated corpora derived out of the texts from multiple intuitions and gold standard annotation are tested on a single institution's rule set has the performances in F1 score of 0.876, 0.706 and 0.694, respectively. The study as a consortium effort of the N3C NLP subgroup demonstrates the feasibility of creating a federated NLP algorithm development and benchmarking platform to enhance multi-institution clinical NLP study.