A desirable property of learning systems is to be both effective and interpretable. Towards this goal, recent models have been proposed that first generate an extractive explanation from the input text and then generate a prediction on just the explanation called explain-then-predict models. These models primarily consider the task input as a supervision signal in learning an extractive explanation and do not effectively integrate rationales data as an additional inductive bias to improve task performance. We propose a novel yet simple approach ExPred, that uses multi-task learning in the explanation generation phase effectively trading-off explanation and prediction losses. And then we use another prediction network on just the extracted explanations for optimizing the task performance. We conduct an extensive evaluation of our approach on three diverse language datasets -- fact verification, sentiment classification, and QA -- and find that we substantially outperform existing approaches.

The development and deployment of machine learning (ML) systems can be executed easily with modern tools, but the process is typically rushed and means-to-an-end. The lack of diligence can lead to technical debt, scope creep and misaligned objectives, model misuse and failures, and expensive consequences. Engineering systems, on the other hand, follow well-defined processes and testing standards to streamline development for high-quality, reliable results. The extreme is spacecraft systems, where mission critical measures and robustness are ingrained in the development process. Drawing on experience in both spacecraft engineering and ML (from research through product across domain areas), we have developed a proven systems engineering approach for machine learning development and deployment. Our "Machine Learning Technology Readiness Levels" (MLTRL) framework defines a principled process to ensure robust, reliable, and responsible systems while being streamlined for ML workflows, including key distinctions from traditional software engineering. Even more, MLTRL defines a lingua franca for people across teams and organizations to work collaboratively on artificial intelligence and machine learning technologies. Here we describe the framework and elucidate it with several real world use-cases of developing ML methods from basic research through productization and deployment, in areas such as medical diagnostics, consumer computer vision, satellite imagery, and particle physics.

In deep learning, models typically reuse the same parameters for all inputs. Mixture of Experts (MoE) defies this and instead selects different parameters for each incoming example. The result is a sparsely-activated model -- with outrageous numbers of parameters -- but a constant computational cost. However, despite several notable successes of MoE, widespread adoption has been hindered by complexity, communication costs and training instability -- we address these with the Switch Transformer. We simplify the MoE routing algorithm and design intuitive improved models with reduced communication and computational costs. Our proposed training techniques help wrangle the instabilities and we show large sparse models may be trained, for the first time, with lower precision (bfloat16) formats. We design models based off T5-Base and T5-Large to obtain up to 7x increases in pre-training speed with the same computational resources. These improvements extend into multilingual settings where we measure gains over the mT5-Base version across all 101 languages. Finally, we advance the current scale of language models by pre-training up to trillion parameter models on the "Colossal Clean Crawled Corpus" and achieve a 4x speedup over the T5-XXL model.

Amid the pandemic COVID-19, the world is facing unprecedented infodemic with the proliferation of both fake and real information. Considering the problematic consequences that the COVID-19 fake-news have brought, the scientific community has put effort to tackle it. To contribute to this fight against the infodemic, we aim to achieve a robust model for the COVID-19 fake-news detection task proposed at CONSTRAINT 2021 (FakeNews-19) by taking two separate approaches: 1) fine-tuning transformers based language models with robust loss functions and 2) removing harmful training instances through influence calculation. We further evaluate the robustness of our models by evaluating on different COVID-19 misinformation test set (Tweets-19) to understand model generalization ability. With the first approach, we achieve 98.13% for weighted F1 score (W-F1) for the shared task, whereas 38.18% W-F1 on the Tweets-19 highest. On the contrary, by performing influence data cleansing, our model with 99% cleansing percentage can achieve 54.33% W-F1 score on Tweets-19 with a trade-off. By evaluating our models on two COVID-19 fake-news test sets, we suggest the importance of model generalization ability in this task to step forward to tackle the COVID-19 fake-news problem in online social media platforms.

With the rapid growth of display devices, quality inspection via machine vision technology has become increasingly important for flat-panel displays (FPD) industry. This paper discloses a novel visual inspection system for liquid crystal display (LCD), which is currently a dominant type in the FPD industry. The system is based on two cornerstones: robust/high-performance defect recognition model and cognitive visual inspection service architecture. A hybrid application of conventional computer vision technique and the latest deep convolutional neural network (DCNN) leads to an integrated defect detection, classfication and impact evaluation model that can be economically trained with only image-level class annotations to achieve a high inspection accuracy. In addition, the properly trained model is robust to the variation of the image qulity, significantly alleviating the dependency between the model prediction performance and the image aquisition environment. This in turn justifies the decoupling of the defect recognition functions from the front-end device to the back-end serivce, motivating the design and realization of the cognitive visual inspection service architecture. Empirical case study is performed on a large-scale real-world LCD dataset from a manufacturing line with different layers and products, which shows the promising utility of our system, which has been deployed in a real-world LCD manufacturing line from a major player in the world.

Accurate and reliable state of charge (SoC) estimation becomes increasingly important to provide a stable and efficient environment for Lithium-ion batteries (LiBs) powered devices. Most data-driven SoC models are built for a fixed ambient temperature, which neglect the high sensitivity of LiBs to temperature and may cause severe prediction errors. Nevertheless, a systematic evaluation of the impact of temperature on SoC estimation and ways for a prompt adjustment of the estimation model to new temperatures using limited data have been hardly discussed. To solve these challenges, a novel SoC estimation method is proposed by exploiting temporal dynamics of measurements and transferring consistent estimation ability among different temperatures. First, temporal dynamics, which is presented by correlations between the past fluctuation and the future motion, is extracted using canonical variate analysis. Next, two models, including a reference SoC estimation model and an estimation ability monitoring model, are developed with temporal dynamics. The monitoring model provides a path to quantitatively evaluate the influences of temperature on SoC estimation ability. After that, once the inability of the reference SoC estimation model is detected, consistent temporal dynamics between temperatures are selected for transfer learning. Finally, the efficacy of the proposed method is verified through a benchmark. Our proposed method not only reduces prediction errors at fixed temperatures (e.g., reduced by 24.35% at -20{\deg}C, 49.82% at 25{\deg}C) but also improves prediction accuracies at new temperatures.

Recent advances in unsupervised domain adaptation (UDA) show that transferable prototypical learning presents a powerful means for class conditional alignment, which encourages the closeness of cross-domain class centroids. However, the cross-domain inner-class compactness and the underlying fine-grained subtype structure remained largely underexplored. In this work, we propose to adaptively carry out the fine-grained subtype-aware alignment by explicitly enforcing the class-wise separation and subtype-wise compactness with intermediate pseudo labels. Our key insight is that the unlabeled subtypes of a class can be divergent to one another with different conditional and label shifts, while inheriting the local proximity within a subtype. The cases of with or without the prior information on subtype numbers are investigated to discover the underlying subtype structure in an online fashion. The proposed subtype-aware dynamic UDA achieves promising results on medical diagnosis tasks.

We introduce the "Incremental Implicitly-Refined Classi-fication (IIRC)" setup, an extension to the class incremental learning setup where the incoming batches of classes have two granularity levels. i.e., each sample could have a high-level (coarse) label like "bear" and a low-level (fine) label like "polar bear". Only one label is provided at a time, and the model has to figure out the other label if it has already learnfed it. This setup is more aligned with real-life scenarios, where a learner usually interacts with the same family of entities multiple times, discovers more granularity about them, while still trying not to forget previous knowledge. Moreover, this setup enables evaluating models for some important lifelong learning challenges that cannot be easily addressed under the existing setups. These challenges can be motivated by the example "if a model was trained on the class bear in one task and on polar bear in another task, will it forget the concept of bear, will it rightfully infer that a polar bear is still a bear? and will it wrongfully associate the label of polar bear to other breeds of bear?". We develop a standardized benchmark that enables evaluating models on the IIRC setup. We evaluate several state-of-the-art lifelong learning algorithms and highlight their strengths and limitations. For example, distillation-based methods perform relatively well but are prone to incorrectly predicting too many labels per image. We hope that the proposed setup, along with the benchmark, would provide a meaningful problem setting to the practitioners

The remarkable ability of deep neural networks to perfectly fit training data when optimized by gradient-based algorithms is yet to be fully explained theoretically. Explanations by recent theoretical works rely on the networks to be wider by orders of magnitude than the ones used in practice. In this work, we take a step towards closing the gap between theory and practice. We show that a randomly initialized deep neural network with ReLU activation converges to a global minimum in a logarithmic number of gradient-descent iterations, under a considerably milder condition on its width. Our analysis is based on a novel technique of training a network with fixed activation patterns. We study the unique properties of the technique that allow an improved convergence, and can be transformed at any time to an equivalent ReLU network of a reasonable size. We derive a tight finite-width Neural Tangent Kernel (NTK) equivalence, suggesting that neural networks trained with our technique generalize well at least as good as its NTK, and it can be used to study generalization as well.

Despite over three hundred years of effort, no solutions exist for predicting when a general planetary configuration will become unstable. We introduce a deep learning architecture to push forward this problem for compact systems. While current machine learning algorithms in this area rely on scientist-derived instability metrics, our new technique learns its own metrics from scratch, enabled by a novel internal structure inspired from dynamics theory. Our Bayesian neural network model can accurately predict not only if, but also when a compact planetary system with three or more planets will go unstable. Our model, trained directly from short N-body time series of raw orbital elements, is more than two orders of magnitude more accurate at predicting instability times than analytical estimators, while also reducing the bias of existing machine learning algorithms by nearly a factor of three. Despite being trained on compact resonant and near-resonant three-planet configurations, the model demonstrates robust generalization to both non-resonant and higher multiplicity configurations, in the latter case outperforming models fit to that specific set of integrations. The model computes instability estimates up to five orders of magnitude faster than a numerical integrator, and unlike previous efforts provides confidence intervals on its predictions. Our inference model is publicly available in the SPOCK package, with training code open-sourced.