We present a self-supervised method to disentangle factors of variation in high-dimensional data that does not rely on prior knowledge of the underlying variation profile (e.g., no assumptions on the number or distribution of the individual latent variables to be extracted). In this method which we call NashAE, high-dimensional feature disentanglement is accomplished in the low-dimensional latent space of a standard autoencoder (AE) by promoting the discrepancy between each encoding element and information of the element recovered from all other encoding elements. Disentanglement is promoted efficiently by framing this as a minmax game between the AE and an ensemble of regression networks which each provide an estimate of an element conditioned on an observation of all other elements. We quantitatively compare our approach with leading disentanglement methods using existing disentanglement metrics. Furthermore, we show that NashAE has increased reliability and increased capacity to capture salient data characteristics in the learned latent representation.
Lighter and faster image restoration (IR) models are crucial for the deployment on resource-limited devices. Binary neural network (BNN), one of the most promising model compression methods, can dramatically reduce the computations and parameters of full-precision convolutional neural networks (CNN). However, there are different properties between BNN and full-precision CNN, and we can hardly use the experience of designing CNN to develop BNN. In this study, we reconsider components in binary convolution, such as residual connection, BatchNorm, activation function, and structure, for IR tasks. We conduct systematic analyses to explain each component's role in binary convolution and discuss the pitfalls. Specifically, we find that residual connection can reduce the information loss caused by binarization; BatchNorm can solve the value range gap between residual connection and binary convolution; The position of the activation function dramatically affects the performance of BNN. Based on our findings and analyses, we design a simple yet efficient basic binary convolution unit (BBCU). Furthermore, we divide IR networks into four parts and specially design variants of BBCU for each part to explore the benefit of binarizing these parts. We conduct experiments on different IR tasks, and our BBCU significantly outperforms other BNNs and lightweight models, which shows that BBCU can serve as a basic unit for binarized IR networks. All codes and models will be released.
Handwritten text recognition for historical documents is an important task but it remains difficult due to a lack of sufficient training data in combination with a large variability of writing styles and degradation of historical documents. While recurrent neural network architectures are commonly used for handwritten text recognition, they are often computationally expensive to train and the benefit of recurrence drastically differs by task. For these reasons, it is important to consider non-recurrent architectures. In the context of handwritten date recognition, we propose an architecture based on the EfficientNetV2 class of models that is fast to train, robust to parameter choices, and accurately transcribes handwritten dates from a number of sources. For training, we introduce a database containing almost 10 million tokens, originating from more than 2.2 million handwritten dates which are segmented from different historical documents. As dates are some of the most common information on historical documents, and with historical archives containing millions of such documents, the efficient and automatic transcription of dates has the potential to lead to significant cost-savings over manual transcription. We show that training on handwritten text with high variability in writing styles result in robust models for general handwritten text recognition and that transfer learning from the DARE system increases transcription accuracy substantially, allowing one to obtain high accuracy even when using a relatively small training sample.
Prediction uncertainty estimation has clinical significance as it can potentially quantify prediction reliability. Clinicians may trust 'blackbox' models more if robust reliability information is available, which may lead to more models being adopted into clinical practice. There are several deep learning-inspired uncertainty estimation techniques, but few are implemented on medical datasets -- fewer on single institutional datasets/models. We sought to compare dropout variational inference (DO), test-time augmentation (TTA), conformal predictions, and single deterministic methods for estimating uncertainty using our model trained to predict feeding tube placement for 271 head and neck cancer patients treated with radiation. We compared the area under the curve (AUC), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) trends for each method at various cutoffs that sought to stratify patients into 'certain' and 'uncertain' cohorts. These cutoffs were obtained by calculating the percentile "uncertainty" within the validation cohort and applied to the testing cohort. Broadly, the AUC, sensitivity, and NPV increased as the predictions were more 'certain' -- i.e., lower uncertainty estimates. However, when a majority vote (implementing 2/3 criteria: DO, TTA, conformal predictions) or a stricter approach (3/3 criteria) were used, AUC, sensitivity, and NPV improved without a notable loss in specificity or PPV. Especially for smaller, single institutional datasets, it may be important to evaluate multiple estimations techniques before incorporating a model into clinical practice.
Paging is a prototypical problem in the area of online algorithms. It has also played a central role in the development of learning-augmented algorithms -- a recent line of research that aims to ameliorate the shortcomings of classical worst-case analysis by giving algorithms access to predictions. Such predictions can typically be generated using a machine learning approach, but they are inherently imperfect. Previous work on learning-augmented paging has investigated predictions on (i) when the current page will be requested again (reoccurrence predictions), (ii) the current state of the cache in an optimal algorithm (state predictions), (iii) all requests until the current page gets requested again, and (iv) the relative order in which pages are requested. We study learning-augmented paging from the new perspective of requiring the least possible amount of predicted information. More specifically, the predictions obtained alongside each page request are limited to one bit only. We consider two natural such setups: (i) discard predictions, in which the predicted bit denotes whether or not it is ``safe'' to evict this page, and (ii) phase predictions, where the bit denotes whether the current page will be requested in the next phase (for an appropriate partitioning of the input into phases). We develop algorithms for each of the two setups that satisfy all three desirable properties of learning-augmented algorithms -- that is, they are consistent, robust and smooth -- despite being limited to a one-bit prediction per request. We also present lower bounds establishing that our algorithms are essentially best possible.
Mobile UI understanding is important for enabling various interaction tasks such as UI automation and accessibility. Previous mobile UI modeling often depends on the view hierarchy information of a screen, which directly provides the structural data of the UI, with the hope to bypass challenging tasks of visual modeling from screen pixels. However, view hierarchy is not always available, and is often corrupted with missing object descriptions or misaligned bounding box positions. As a result, although using view hierarchy offers some short-term gains, it may ultimately hinder the applicability and performance of the model. In this paper, we propose Spotlight, a vision-only approach for mobile UI understanding. Specifically, we enhance a vision-language model that only takes the screenshot of the UI and a region of interest on the screen -- the focus -- as the input. This general architecture is easily scalable and capable of performing a range of UI modeling tasks. Our experiments show that our model obtains SoTA results on several representative UI tasks and outperforms previous methods that use both screenshots and view hierarchies as input. Furthermore, we explore the multi-task learning and few-shot prompting capacity of the proposed models, demonstrating promising results in the multi-task learning direction.
Domain generalization (DG), aiming to make models work on unseen domains, is a surefire way toward general artificial intelligence. Limited by the scale and diversity of current DG datasets, it is difficult for existing methods to scale to diverse domains in open-world scenarios (e.g., science fiction and pixelate style). Therefore, the source-free domain generalization (SFDG) task is necessary and challenging. To address this issue, we propose an approach based on large-scale vision-language pretraining models (e.g., CLIP), which exploits the extensive domain information embedded in it. The proposed scheme generates diverse prompts from a domain bank that contains many more diverse domains than existing DG datasets. Furthermore, our method yields domain-unified representations from these prompts, thus being able to cope with samples from open-world domains. Extensive experiments on mainstream DG datasets, namely PACS, VLCS, OfficeHome, and DomainNet, show that the proposed method achieves competitive performance compared to state-of-the-art (SOTA) DG methods that require source domain data for training. Besides, we collect a small datasets consists of two domains to evaluate the open-world domain generalization ability of the proposed method. The source code and the dataset will be made publicly available at https://github.com/muse1998/Source-Free-Domain-Generalization
Traditional automated metrics for evaluating conditional natural language generation use pairwise comparisons between a single generated text and the best-matching gold-standard ground truth text. When multiple ground truths are available, scores are aggregated using an average or max operation across references. While this approach works well when diversity in the ground truth data (i.e. dispersion of the distribution of conditional texts) can be ascribed to noise, such as in automated speech recognition, it does not allow for robust evaluation in the case where diversity in the ground truths represents signal for the model. In this work we argue that existing metrics are not appropriate for domains such as visual description or summarization where ground truths are semantically diverse, and where the diversity in those captions captures useful additional information about the context. We propose a novel paradigm for multi-candidate evaluation of conditional language generation models, and a new family of metrics that compare the distributions of reference and model-generated caption sets using small sample sets of each. We demonstrate the utility of our approach with a case study in visual description: where we show that existing models optimize for single-description quality over diversity, and gain some insights into how sampling methods and temperature impact description quality and diversity.
This paper presents MOAT, a family of neural networks that build on top of MObile convolution (i.e., inverted residual blocks) and ATtention. Unlike the current works that stack separate mobile convolution and transformer blocks, we effectively merge them into a MOAT block. Starting with a standard Transformer block, we replace its multi-layer perceptron with a mobile convolution block, and further reorder it before the self-attention operation. The mobile convolution block not only enhances the network representation capacity, but also produces better downsampled features. Our conceptually simple MOAT networks are surprisingly effective, achieving 89.1% top-1 accuracy on ImageNet-1K with ImageNet-22K pretraining. Additionally, MOAT can be seamlessly applied to downstream tasks that require large resolution inputs by simply converting the global attention to window attention. Thanks to the mobile convolution that effectively exchanges local information between pixels (and thus cross-windows), MOAT does not need the extra window-shifting mechanism. As a result, on COCO object detection, MOAT achieves 59.2% box AP with 227M model parameters (single-scale inference, and hard NMS), and on ADE20K semantic segmentation, MOAT attains 57.6% mIoU with 496M model parameters (single-scale inference). Finally, the tiny-MOAT family, obtained by simply reducing the channel sizes, also surprisingly outperforms several mobile-specific transformer-based models on ImageNet. We hope our simple yet effective MOAT will inspire more seamless integration of convolution and self-attention. Code is made publicly available.
Data-driven decision making is becoming an integral part of manufacturing companies. Data is collected and commonly used to improve efficiency and produce high quality items for the customers. IoT-based and other forms of object tracking are an emerging tool for collecting movement data of objects/entities (e.g. human workers, moving vehicles, trolleys etc.) over space and time. Movement data can provide valuable insights like process bottlenecks, resource utilization, effective working time etc. that can be used for decision making and improving efficiency. Turning movement data into valuable information for industrial management and decision making requires analysis methods. We refer to this process as movement analytics. The purpose of this document is to review the current state of work for movement analytics both in manufacturing and more broadly. We survey relevant work from both a theoretical perspective and an application perspective. From the theoretical perspective, we put an emphasis on useful methods from two research areas: machine learning, and logic-based knowledge representation. We also review their combinations in view of movement analytics, and we discuss promising areas for future development and application. Furthermore, we touch on constraint optimization. From an application perspective, we review applications of these methods to movement analytics in a general sense and across various industries. We also describe currently available commercial off-the-shelf products for tracking in manufacturing, and we overview main concepts of digital twins and their applications.