Regularizing Nighttime Weirdness: Efficient Self-supervised Monocular Depth Estimation in the Dark

Monocular depth estimation aims at predicting depth from a single image or video. Recently, self-supervised methods draw much attention since they are free of depth annotations and achieve impressive performance on several daytime benchmarks. However, they produce weird outputs in more challenging nighttime scenarios because of low visibility and varying illuminations, which bring weak textures and break brightness-consistency assumption, respectively. To address these problems, in this paper we propose a novel framework with several improvements: (1) we introduce Priors-Based Regularization to learn distribution knowledge from unpaired depth maps and prevent model from being incorrectly trained; (2) we leverage Mapping-Consistent Image Enhancement module to enhance image visibility and contrast while maintaining brightness consistency; and (3) we present Statistics-Based Mask strategy to tune the number of removed pixels within textureless regions, using dynamic statistics. Experimental results demonstrate the effectiveness of each component. Meanwhile, our framework achieves remarkable improvements and state-of-the-art results on two nighttime datasets.

Learning to Aggregate and Personalize 3D Face from In-the-Wild Photo Collection

Non-parametric face modeling aims to reconstruct 3D face only from images without shape assumptions. While plausible facial details are predicted, the models tend to over-depend on local color appearance and suffer from ambiguous noise. To address such problem, this paper presents a novel Learning to Aggregate and Personalize (LAP) framework for unsupervised robust 3D face modeling. Instead of using controlled environment, the proposed method implicitly disentangles ID-consistent and scene-specific face from unconstrained photo set. Specifically, to learn ID-consistent face, LAP adaptively aggregates intrinsic face factors of an identity based on a novel curriculum learning approach with relaxed consistency loss. To adapt the face for a personalized scene, we propose a novel attribute-refining network to modify ID-consistent face with target attribute and details. Based on the proposed method, we make unsupervised 3D face modeling benefit from meaningful image facial structure and possibly higher resolutions. Extensive experiments on benchmarks show LAP recovers superior or competitive face shape and texture, compared with state-of-the-art (SOTA) methods with or without prior and supervision.

BNN - BN = ?: Training Binary Neural Networks without Batch Normalization

Batch normalization (BN) is a key facilitator and considered essential for state-of-the-art binary neural networks (BNN). However, the BN layer is costly to calculate and is typically implemented with non-binary parameters, leaving a hurdle for the efficient implementation of BNN training. It also introduces undesirable dependence between samples within each batch. Inspired by the latest advance on Batch Normalization Free (BN-Free) training, we extend their framework to training BNNs, and for the first time demonstrate that BNs can be completed removed from BNN training and inference regimes. By plugging in and customizing techniques including adaptive gradient clipping, scale weight standardization, and specialized bottleneck block, a BN-free BNN is capable of maintaining competitive accuracy compared to its BN-based counterpart. Extensive experiments validate the effectiveness of our proposal across diverse BNN backbones and datasets. For example, after removing BNs from the state-of-the-art ReActNets, it can still be trained with our proposed methodology to achieve 92.08%, 68.34%, and 68.0% accuracy on CIFAR-10, CIFAR-100, and ImageNet respectively, with marginal performance drop (0.23%~0.44% on CIFAR and 1.40% on ImageNet). Codes and pre-trained models are available at: https://github.com/VITA-Group/BNN_NoBN.

Formalism for Supporting the Development of Verifiably Safe Medical Guidelines with Statecharts

Improving the effectiveness and safety of patient care is the ultimate objective for medical cyber-physical systems. Many medical best practice guidelines exist, but most of the existing guidelines in handbooks are difficult for medical staff to remember and apply clinically. Furthermore, although the guidelines have gone through clinical validations, validations by medical professionals alone do not provide guarantees for the safety of medical cyber-physical systems. Hence, formal verification is also needed. The paper presents the formal semantics for a framework that we developed to support the development of verifiably safe medical guidelines. The framework allows computer scientists to work together with medical professionals to transform medical best practice guidelines into executable statechart models, Yakindu in particular, so that medical functionalities and properties can be quickly prototyped and validated. Existing formal verification technologies, UPPAAL timed automata in particular, is integrated into the framework to provide formal verification capabilities to verify safety properties. However, some components used/built into the framework, such as the open-source Yakindu statecharts as well as the transformation rules from statecharts to timed automata, do not have built-in semantics. The ambiguity becomes unavoidable unless formal semantics is defined for the framework, which is what the paper is to present.

Joint Extraction of Entities and Relations Based on a Novel Decomposition Strategy

Joint extraction of entities and relations aims to detect entity pairs along with their relations using a single model. Prior work typically solves this task in the extract-then-classify or unified labeling manner. However, these methods either suffer from the redundant entity pairs, or ignore the important inner structure in the process of extracting entities and relations. To address these limitations, in this paper, we first decompose the joint extraction task into two interrelated subtasks, namely HE extraction and TER extraction. The former subtask is to distinguish all head-entities that may be involved with target relations, and the latter is to identify corresponding tail-entities and relations for each extracted head-entity. Next, these two subtasks are further deconstructed into several sequence labeling problems based on our proposed span-based tagging scheme, which are conveniently solved by a hierarchical boundary tagger and a multi-span decoding algorithm. Owing to the reasonable decomposition strategy, our model can fully capture the semantic interdependency between different steps, as well as reduce noise from irrelevant entity pairs. Experimental results show that our method outperforms previous work by 5.2%, 5.9% and 21.5% (F1 score), achieving a new state-of-the-art on three public datasets

Non-parametric dark energy reconstruction using the tomographic Alcock-Paczynski test

The tomographic Alcock-Paczynski (AP) method can result in tight cosmological constraints by using small and intermediate clustering scales of the large scale structure (LSS) of the galaxy distribution. By focusing on the redshift dependence, the AP distortion can be distinguished from the distortions produced by the redshift space distortions (RSD). In this work, we combine the tomographic AP method with other recent observational datasets of SNIa+BAO+CMB+$H_0$ to reconstruct the dark energy equation-of-state $w$ in a non-parametric form. The result favors a dynamical DE at $zlesssim1$, and shows a mild deviation ($lesssim2sigma$) from $w=-1$ at $z=0.5-0.7$. We find the addition of the AP method improves the low redshift ($zlesssim0.7$) constraint by $sim50%$.

Model and Integrate Medical Resource Availability into Verifiably Correct Executable Medical Guidelines - Technical Report

Improving effectiveness and safety of patient care is an ultimate objective for medical cyber-physical systems. A recent study shows that the patients death rate can be reduced by computerizing medical guidelines. Most existing medical guideline models are validated and/or verified based on the assumption that all necessary medical resources needed for a patient care are always available. However, the reality is that some medical resources, such as special medical equipment or medical specialists, can be temporarily unavailable for an individual patient. In such cases, safety properties validated and/or verified in existing medical guideline models without considering medical resource availability may not hold any more. The paper argues that considering medical resource availability is essential in building verifiably correct executable medical guidelines. We present an approach to explicitly and separately model medical resource availability and automatically integrate resource availability models into an existing statechart-based computerized medical guideline model. This approach requires minimal change in existing medical guideline models to take into consideration of medical resource availability in validating and verifying medical guideline models. A simplified stroke scenario is used as a case study to investigate the effectiveness and validity of our approach.

Model and Integrate Medical Resource Available Times and Relationships in Verifiably Correct Executable Medical Best Practice Guideline Models (Extended Version)

Improving patient care safety is an ultimate objective for medical cyber-physical systems. A recent study shows that the patients death rate is significantly reduced by computerizing medical best practice guidelines. Recent data also show that some morbidity and mortality in emergency care are directly caused by delayed or interrupted treatment due to lack of medical resources. However, medical guidelines usually do not provide guidance on medical resource demands and how to manage potential unexpected delays in resource availability. If medical resources are temporarily unavailable, safety properties in existing executable medical guideline models may fail which may cause increased risk to patients under care. The paper presents a separately model and jointly verify (SMJV) architecture to separately model medical resource available times and relationships and jointly verify safety properties of existing medical best practice guideline models with resource models being integrated in. The SMJV architecture allows medical staff to effectively manage medical resource demands and unexpected resource availability delays during emergency care. The separated modeling approach also allows different domain professionals to make independent model modifications, facilitates the management of frequent resource availability changes, and enables resource statechart reuse in multiple medical guideline models. A simplified stroke scenario is used as a case study to investigate the effectiveness and validity of the SMJV architecture. The case study indicates that the SMJV architecture is able to identify unsafe properties caused by unexpected resource delays.

Design Verifiably Correct Model Patterns to Facilitate Modeling Medical Best Practice Guidelines with Statecharts (Technical Report)

Improving patient care safety is an ultimate objective for medical cyber-physical systems. A recent study shows that the patients death rate can be significantly reduced by computerizing medical best practice guidelines. To facilitate the development of computerized medical best practice guidelines, statecharts are often used as a modeling tool because of their high resemblances to disease and treatment models and their capabilities to provide rapid prototyping and simulation for clinical validations. However, some implementations of statecharts, such as Yakindu statecharts, are priority-based and have synchronous execution semantics which makes it difficult to model certain functionalities that are essential in modeling medical guidelines, such as two-way communications and configurable execution orders. Rather than introducing new statechart elements or changing the statechart implementations underline semantics, we use existing basic statechart elements to design model patterns for the commonly occurring issues. In particular, we show the design of model patterns for two-way communications and configurable execution orders and formally prove the correctness of these model patterns. We further use a simplified airway laser surgery scenario as a case study to demonstrate how the developed model patterns address the two-way communication and configurable execution order issues and their impact on validation and verification of medical safety properties.