Limited Supervision Research Articles

Implementasi Kebijakan Larangan Parkir Liar di Kota Surabaya (Studi Kasus Parkir Pasar Kapasan)

The Kapasan Market area is one of the areas in the city of Surabaya where parking problems often occur. Some illegal parking attendants often park visitors' vehicles on the side of the road because of limited parking spaces and supervision, causing traffic jams around Kapasan Market which the public often complains about. The Surabaya City Transportation Department (Dishub) is guided by Surabaya City Regional Regulation Number 3 of 2018 concerning the Implementation of Parking in the City of Surabaya in implementing the parking ban policy in the City of Surabaya, including in the Kapasan Market area. This research aims to find out whether the parking ban policy has been implemented well and is appropriate, especially in the Kapasan Market area. This research uses a qualitative approach with descriptive methods. The research uses primary data obtained through direct observation of research objects, interviews, and documentation as well as secondary data from documents, books, the internet, and other written sources. Determining informants for interview activities using purposive sampling techniques to the Head of the Public Roadside Parking Division. The research results show that the implementation of the policy of prohibiting illegal parking in the city of Surabaya, especially at Kapasan Market, is quite good. four of the six policy implementation indicators according to Van Meter and Van Horn have been met so that the public can feel the response from the government regarding illegal parking that they complain about.

Teacher-Student Prototype Enhancement Network for a Few-Shot Remote Sensing Scene Classification

Few-shot remote sensing scene classification identifies new classes from limited labeled samples where the great challenges are intraclass diversity, interclass similarity, and limited supervision. To alleviate these problems, a teacher-student prototype enhancement network is proposed for a few-shot remote sensing scene classification. Instead of introducing an attentional mechanism in mainstream studies, a prototype enhancement module is recommended to adaptively select high-confidence query samples, which can enhance the support prototype representations to emphasize intraclass and interclass relationships. The construction of a few-shot teacher model generates more discriminative predictive representations with inputs from many labeled samples, thus providing a strong supervisory signal to the student model and encouraging the network to achieve accurate classification with a limited number of labeled samples. Extensive experiments of four public datasets, including NWPU-remote sens ing image scene classification (NWPU-RESISC45), aerial image dataset (AID), UC Merced, and WHU-RS19, demonstrate that this method achieves superior competitive performance than the state-of-the-art methods on five-way, one-shot, and five-shot classifications.

Curriculum Development for the South African Essential Steps in Managing Obstetric Emergencies (ESMOE) Anesthesiology Training Module: A Delphi Study.

The United Nations Sustainable Development Goal 3 (SDG3) for 2030 aims at <70 maternal deaths per 100,000 live births. South Africa (SA) falls short of this goal and most deaths occur in district and regional hospitals. Due to low anesthesiologist (specialist anesthetist) numbers in the public sector, the anesthetic workforce in these hospitals consists mainly of nonspecialist (general practitioner) junior doctors with limited supervision. The Essential Steps in Managing Obstetric Emergencies (ESMOE) training program for interns was introduced in 2008 to reduce maternal deaths in SA. Training is not consistently offered at all intern-training institutions and it has not been recently revised. This study sought expert suggestions to guide a revision of the regulation, pedagogy, and content of the ESMOE anesthetic module. A 3-round consensus-seeking modified Delphi technique was used whereby experts rated current and suggested learning outcomes as well as pedagogical- and regulatory aspects of the ESMOE anesthetic module in an anonymous, online, Likert-scale questionnaire. Consensus was defined as ≥70% agreement. Open-ended questions sought motivations for decisions and further suggestions for program improvement. Between March 4, 2021 and August 18, 2021, a total of 17/24 (71%), 16/17 (94%), and 15/16 (94%) consenting experts completed the first, second, and third rounds, respectively. During the 3 rounds, the panel achieved consensus on all but 2 questionnaire items. The panel agreed that ESMOE anesthesia training should be mandatory and that it should be offered through a blended learning model. They further concurred that the current learning outcomes should be retained, and suggested additional outcomes based on the most common causes of maternal mortality in SA. Attendance of the ESMOE anesthesia module should be mandatory. Its pedagogy should be updated to a blended learning style to benefit the current digital native intern generation. The content should be updated to address the main causes of maternal mortality in SA in line with SDG3. A national obstetric anesthesia guideline should be considered, and lea-dership and collaboration are required to improve the alignment of undergraduate, internship, and junior doctor anesthesia training in SA. The content, process, and stakeholder engagement suggestions in this study can assist short-course-based anesthesia workforce training in similar global contexts.

Dealing with partial labels by knowledge distillation

Partial label learning (PLL) is a weakly supervised methodology dealing with tasks that have annotation problems by replacing the single label with a collection of candidate labels. Compared to single labels, utilizing partial labels faces challenges: (1) The limited supervision and sensitivity to the false positive candidates; (2) Situations where the ground truth is not in the candidate label sets (noisy PLs). However, in the case that there exists a subset of samples that can be easily labeled (referred to as clean samples), the existing PLL paradigm needlessly assigns these instances with partial labels randomly. To better utilize the clean samples, and alleviate the obstacles of adopting partial labels, we proposed a specific Partial Label Knowledge Distillation (PLKD) framework to distill the knowledge from the samples with low annotating cost, further guiding partial label samples with limited supervision in these scenarios. The teacher model of PLKD was pre-trained on the clean samples with a single label, which can reduce the effect of noisy PLs when training on the remaining PLL samples. Additionally, recognizing that the existing candidate labels are sampled under the uniform distribution, which may not reflect real-life scenarios, we also proposed a label-specific candidate generation method. Correspondingly, a new loss function based on our generation method is presented to evaluate the distinction between partial labels and predictions. Furthermore, we also present a partial-label guided version, denoted as PLKD-pl, to alleviate the teacher’s risk of over-confidence when the distribution between the clean set and partial label set varies widely. Extensive experimental evaluations have been conducted to demonstrate the superiority of PLKD over six state-of-the-art counterparts.

Determining calf traceability and cow–calf relationships in extensive farming using geolocation collars and BLE ear tags

Extensive farming is often considered very beneficial to animals for its potential to enhance animal welfare, providing animals with free-range access to their natural environment where they can engage in innate behaviors like grazing and exploration. However, despite these benefits, extensive production still faces welfare and health challenges due to unpredictable weather conditions and limited supervision by stockholders. Moreover, increasing consumer demand for information regarding food quality, safety, and production conditions poses a challenge for extensive farming, where animals are less controlled. Precision livestock farming (PLF) emerges as a possible solution by enabling the continuous real-time monitoring of the health, welfare, and behavior of animals. A novel approach combining geolocation collars for cows and Bluetooth low energy (BLE) ear tags for calves appears promising to enhance traceability and monitoring in extensive farming. Nevertheless, challenges persist, including limitations in the data transmission capacity and associated costs. This study evaluated the effectiveness of combining geolocation collars and BLE ear tags for monitoring calf traceability and cow–calf relationships across three scenarios: 1) Equilibrated: same collar/ear tag ratio; good coverage; 2) intermediate: more collars than ear tags; fair coverage; 3) worst: more ear tags than collars; lousy coverage. Our results indicate that the equilibrated scenario (ES) with an equal number of geolocation collars and BLE ear tags, was the best scenario, demonstrating the highest fix rate (22%) and the longest mean consecutive days of detecting the same ear tag (22.30 days), followed by the intermediate scenario (IS) and the worst scenario. In all scenarios, there was a mean period of 14–15 d without detecting a calf. However, this shortcoming can be overcome as calves usually graze alongside their mothers, ensuring comprehensive traceability in farm breeding. Additionally, by comparing differences in the number of ear tags received from offspring compared to other calves, the BLE ear tags successfully associated every mother with their calf in the ES and IS. Finally, this valuable information, would enable the development of a traceability system that ensures transparency and reliability throughout the supply chain and could allow consumers to access to product information related to animal welfare.

Knowledge Graph Alignment Under Scarce Supervision: A General Framework With Active Cross-View Contrastive Learning.

Over recent years, a number of knowledge graphs (KGs) have emerged. Nevertheless, a KG can never reach full completeness. A viable approach to increase the coverage of a KG is KG alignment (KGA). The majority of previous efforts merely focus on the matching between entities, while largely neglect relations. Besides, they heavily rely on labeled data, which are difficult to obtain in practice. To address these issues, in this work, we put forward a general framework to simultaneously align entities and relations under scarce supervision. Our proposal consists of two main components, relation-enhanced active instance selection (RAS), and cross-view contrastive learning (CCL). RAS aims to select the most valuable instances to be labeled with the guidance of relations, while CCL contrasts cross-view representations to augment scarce supervision signals. Our proposal is agnostic to the underlying entity and relation alignment models, and can be used to improve their performance under limited supervision. We conduct experiments on a wide range of popular KG pairs, and the results demonstrate that our proposed model and its components can consistently boost the alignment performance under scarce supervision.

Constrains for attending District Early Intervention Centre (DEIC) of Purba Bardhaman district, West Bengal: Perspectives of both service providers and caregivers of the beneficiaries, a Mixed-Method Study

ABSTRACT Rashtriya Bal Swasthya Karyakram (RBSK) is an ongoing screening program that deals with early detection and management of the 4Ds prevalent in children (defects at birth, diseases, deficiencies, and developmental delays). It runs a District Early Intervention Centre (DEIC), which is an interdisciplinary approach of a multidisciplinary team placed under one roof. It provides referral support to children detected with health conditions during health screening. Aims: This study was conducted to assess the clinicosocial profile characteristics of the beneficiaries attending DEIC and to explore the challenges faced by both the caregivers and service providers. Settings and Design: A facility-based, mixed-method study with phenomenological approach was conducted in Burdwan Medical College of Purba Bardhaman district, West Bengal, from January 2023 to April 2023. Methods and Materials: For quantitative part, record analysis was conducted involving all children enrolled at DEIC from January 2022 to December 2022, and for qualitative part, caregivers and service providers were selected purposively and interviewed using the in-depth-interview (IDI) guide. Inductive thematic analysis was conducted simultaneously till data saturation was achieved. Statistical Analysis Used: The principles of descriptive statistics were applied to organize and present the quantitative data in tables and diagrams. Qualitative data were transcribed into verbatim and analyzed thematically with inductive approach. Results: Out of 3332 children enrolled, 58.4% were males, and 44.2% were aged between 0 and 1 month. 59.3% of them had developmental delays, 7.9% had defects, 42.9% had diseases, and 21.3% had some deficiencies. Major hindering factors in service provision among service providers were resource deficiency, inadequate capacity building, administrative bottlenecks, limited supportive supervision, inadequate advocacy, and awareness deficit. No major constraints were found from caregivers’ perspective in accessing healthcare. Conclusions: DEIC is running with shortage of proper infrastructure and adequate supervision. Periodic training, appropriate awareness building, and continuous flow of logistics are needed to ensure a better quality of life among beneficiaries.

A Cognitive Load Theory Perspective of the Undergraduate Anesthesia Curricula in South Africa.

Safe anesthesia is indispensable to achieve global safe surgery and equitable health care access. The disease burden and lack of specialists in South Africa (SA) require junior, nonspecialist doctors to be fit-for-purpose from day 1 when they provide anesthetic services in peripheral hospitals with limited supervision. Graduating students report low self-perceived preparedness for administering anesthesia, but it is not known how their curricular experiences influence their learning. Cognitive load theory defines intrinsic, extraneous, and germane cognitive loads (subtypes). Intrinsic load relates to learning tasks, extraneous load to distractions, and germane load to students' learning processes. This study used a cognitive load theory lens to explore SA students' experiences of their undergraduate anesthesia training. In a constructivist cross-sectional descriptive study, we explored the qualitative factors that influenced students' curricular experience of undergraduate anesthesia training in SA. Two investigators analyzed the data independently in an initial coding round. An emerging theme of lack of time to achieve the expected outcomes, prompted the use of cognitive load theory as a conceptual framework for further analysis by the 3 authors. The subsequent analysis informed the development and refinement of a final cognitive load theory framework for anesthesia training, the COLOAD (COgnitive LOad in Anesthesia eDucation) framework. Data were collected between November 2017 and February 2019. The 1336 respondents (79% participation) reported a variety of determinants of learning pertaining to all 3 cognitive load subtypes. Participants were novices in an inherently complex environment and experienced a high cognitive load during anesthesia training. The number-, complexity-, and interactivity of tasks influenced intrinsic load, while extraneous load was affected by ineffective instructional methods, external- and internal distractors. Program design, metacognition, and learner motivation impacted germane load. Cognitive load theory provided a useful theoretical basis for understanding students' curricular experiences. The COLOAD framework suggests a microlevel interrelatedness of the constituting elements of the 3 cognitive load subtypes. This has implications for curriculum design, pedagogy, and student support. Learning outcomes development and curriculum mapping are important to ensure a lean curriculum, but measures to enhance germane cognitive load might be equally important to achieve competence. Attention to the hidden curriculum and active promotion of reflective practice might reduce cognitive load in complex learning environments such as anesthesia training.

IDSSI: Image Deturbulence with Semantic and Spatial–Temporal Information

The propagation of light through turbulent media results in refractive index fluctuations, causing distortion and blurring of images captured by imaging equipment. Consequently, it is of paramount importance to mitigate turbulence effects. Deep convolutional neural networks incorporating attention mechanisms have demonstrated remarkable success in dynamic video restoration. However, in most networks, the attention mechanism is limited to capturing simple contextual features, failing to adequately exploit multi-level image features. This paper introduces IDSSI, an effective model that utilizes semantic and spatio–temporal information for turbulence mitigation. A novel two-branch feature extraction structure is proposed, capable of extracting multi-scale enhanced features with semantic information under limited supervision. These perceived semantic features are subsequently fused into global features, enhancing multi-scale perception for turbulence repair. Furthermore, a new Spatial–Temporal feature learning strategy is proposed. This approach facilitates the extraction and modulation of temporal information by obtaining edge cues, effectively concatenating and merging with spatial features. This strategy serves as an efficient alternative to 3D convolution. Experimental results on relevant datasets demonstrate the superiority of the proposed model over current state-of-the-art turbulence repair methods.

ASAS PENGAYOMAN DALAM SISTEM JAMINAN PRODUK HALAL DI KABUPATEN ACEH BARAT

Aceh Qanun No. 8/2016 on Halal Product Guarantee System is based on the principle of protection. The principle of protection is a function of the Government of Aceh to protect the lives of all people, not only Muslims, but all other religious communities. In this case, the Government of Aceh tries to provide a guarantee of legal certainty for products consumed by the community. Based on the data found, there are 94 MSMEs in West Aceh Regency, but only 30 MSMEs have halal certificates. Based on this, there are two problems, namely the implementation of the principle of protection in the halal product guarantee system and how the form of Local Government supervision of the halal product guarantee system in West Aceh Regency. The research method used is empirical juridical which produces descriptive data regarding the Principle of Protection in the Halal Product Guarantee System in West Aceh Regency. The result of the research found that the implementation of the principle of penayoman by the Government of Aceh to provide legal certainty guarantees for products consumed by the public has not been maximized. This can be seen from the number of Micro, Small and Medium Enterprises that do not have halal certificates that are still many. The application of the principle of protection that has not been maximised is also influenced by the supervision system that has not been maximised either. Supervision involves several institutions in the West Aceh District Government such as the Ministry of Religion and the Health Office. The main obstacle that occurs is the lack of funds allocated for this limited supervision so the supervision of halal products for Micro, Small and Medium Enterprises in West Aceh is not optimal

Pengawasan Kedisiplinan Peserta Didik di SMA Sriguna Palembang

This study aims to understand the process of monitoring student discipline at Sriguna High School in Palembang. This research is qualitative with a descriptive approach. The data collection methods used are interviews, observations, and documentation. The data analysis techniques include data reduction, data presentation, and verification and decision-making. To ensure the validity of the data, triangulation is used, which involves time, method, and source. The results of the study indicate that the process of monitoring student discipline at Sriguna High School in Palembang has been carried out well. This includes monitoring to ensure that students arrive on time, do not skip school, monitor student activities within the school environment, and behave according to established regulations. The supporting factors include supporting school activity plans and good cooperation among educators. The inhibiting factors are the lack of cooperation between parents and students and limited teacher supervision of students. In this study, the author concludes that supervision, monitoring, inspection, guidance and direction, disciplinary actions, and corrective actions are necessary for the student discipline monitoring process to be effective and efficient

Estimation of Soil Characteristic Parameters for Electric Mountain Tractor Based on Gauss–Newton Iteration Method

Future field work tasks will require mountain tractors to pass through rough terrain with limited human supervision. The wheel–soil interaction plays a critical role in rugged terrain mobility. In this paper, an algorithm for the estimation of soil characteristic parameters based on the Simpson numerical integration method and Gauss–Newton iteration method is presented. These parameters can be used for passability prediction or in a traction control algorithm to improve tractor mobility and to plan safe operation paths for autonomous navigation systems. To verify the effectiveness of the solving algorithm, different initial values and soils were selected for simulation calculations of soil characteristic parameters such as internal friction angle, settlement index, and the joint parameter of soil cohesion modulus and friction modulus. The results show that the error was kept within 2%, and the calculation time did not exceed 0.84 s, demonstrating high robustness and real-time performance. To test the applicability of the algorithm model, further research was conducted using different wheel parameters of electric mountain tractors under wet clay conditions. The results show that these parameters also have high accuracy and stability with only a few iterations. Thus, the estimation algorithm can meet the requirements of quickly and accurately identifying soil characteristic parameters during tractor operation. A criterion for the passability of wheeled tractors through unknown terrain is proposed, utilizing identified soil parameters.

Semi-Supervised Learning for Multi-Label Cardiovascular Diseases Prediction: A Multi-Dataset Study.

Electrocardiography (ECG) is a non-invasive tool for predicting cardiovascular diseases (CVDs). Current ECG-based diagnosis systems show promising performance owing to the rapid development of deep learning techniques. However, the label scarcity problem, the co-occurrence of multiple CVDs and the poor performance on unseen datasets greatly hinder the widespread application of deep learning-based models. Addressing them in a unified framework remains a significant challenge. To this end, we propose a multi-label semi-supervised model (ECGMatch) to recognize multiple CVDs simultaneously with limited supervision. In the ECGMatch, an ECGAugment module is developed for weak and strong ECG data augmentation, which generates diverse samples for model training. Subsequently, a hyperparameter-efficient framework with neighbor agreement modeling and knowledge distillation is designed for pseudo-label generation and refinement, which mitigates the label scarcity problem. Finally, a label correlation alignment module is proposed to capture the co-occurrence information of different CVDs within labeled samples and propagate this information to unlabeled samples. Extensive experiments on four datasets and three protocols demonstrate the effectiveness and stability of the proposed model, especially on unseen datasets. As such, this model can pave the way for diagnostic systems that achieve robust performance on multi-label CVDs prediction with limited supervision.

Language model-based labeling of German thoracic radiology reports.

The aim of this study was to explore the potential of weak supervision in a deep learning-based label prediction model. The goal was to use this model to extract labels from German free-text thoracic radiology reports on chest X-ray images and for training chest X-ray classification models.The proposed label extraction model for German thoracic radiology reports uses a German BERT encoder as a backbone and classifies a report based on the CheXpert labels. For investigating the efficient use of manually annotated data, the model was trained using manual annotations, weak rule-based labels, and both. Rule-based labels were extracted from 66071 retrospectively collected radiology reports from 2017-2021 (DS 0), and 1091 reports from 2020-2021 (DS 1) were manually labeled according to the CheXpert classes. Label extraction performance was evaluated with respect to mention extraction, negation detection, and uncertainty detection by measuring F1 scores. The influence of the label extraction method on chest X-ray classification was evaluated on a pneumothorax data set (DS 2) containing 6434 chest radiographs with associated reports and expert diagnoses of pneumothorax. For this, DenseNet-121 models trained on manual annotations, rule-based and deep learning-based label predictions, and publicly available data were compared.The proposed deep learning-based labeler (DL) performed on average considerably stronger than the rule-based labeler (RB) for all three tasks on DS 1 with F1 scores of 0.938 vs. 0.844 for mention extraction, 0.891 vs. 0.821 for negation detection, and 0.624 vs. 0.518 for uncertainty detection. Pre-training on DS 0 and fine-tuning on DS 1 performed better than only training on either DS 0 or DS 1. Chest X-ray pneumothorax classification results (DS 2) were highest when trained with DL labels with an area under the receiver operating curve (AUC) of 0.939 compared to RB labels with an AUC of 0.858. Training with manual labels performed slightly worse than training with DL labels with an AUC of 0.934. In contrast, training with a public data set resulted in an AUC of 0.720.Our results show that leveraging a rule-based report labeler for weak supervision leads to improved labeling performance. The pneumothorax classification results demonstrate that our proposed deep learning-based labeler can serve as a substitute for manual labeling requiring only 1000 manually annotated reports for training. · The proposed deep learning-based label extraction model for German thoracic radiology reports performs better than the rule-based model.. · Training with limited supervision outperformed training with a small manually labeled data set.. · Using predicted labels for pneumothorax classification from chest radiographs performed equally to using manual annotations.. Wollek A, Haitzer P, Sedlmeyr T et al. Language modelbased labeling of German thoracic radiology reports. Fortschr Röntgenstr 2024; DOI 10.1055/a-2287-5054.

Assessing sustainable waste management practices in Rajshahi City Corporation: an analysis for local government enhancement using IoT, AI, and Android technology.

This study addresses the pressing challenge of inefficient waste management practices within the Rajshahi City Corporation (RCC), Bangladesh. Despite rapid urbanization and escalating waste generation rates, RCC struggles with diverse waste disposal practices, limited supervision, irregular waste collection schedules, and inadequate disposal infrastructure. In this context, this study examines the possible improvements that could be made by combining the Internet of Things (IoT), artificial intelligence (AI), and Android application to improve waste management methods in the RCC. The study's foundation is a vast amount of information gathered from residents, with particular attention paid to waste disposal methods, the role of the local government, the frequency of waste collection, and public attitudes toward waste management. The results point to a complicated waste management environment with a range of waste disposal practices, little supervision, irregular waste collection, and insufficient disposal methods. The importance of RCC in waste management is emphasized, highlighting the need for proactive measures including effective monitoring, constant waste collection, and routine drain cleaning. Additionally, it is suggested that combining IoT, AI, and Android technology is a possible way to improve waste management procedures. These technologies have the potential to increase productivity, lessen their negative effects on the environment, and produce cleaner, more sustainable urban environments.

Liberating Seen Classes: Boosting Few-Shot and Zero-Shot Text Classification via Anchor Generation and Classification Reframing

Few-shot and zero-shot text classification aim to recognize samples from novel classes with limited labeled samples or no labeled samples at all. While prevailing methods have shown promising performance via transferring knowledge from seen classes to unseen classes, they are still limited by (1) Inherent dissimilarities among classes make the transformation of features learned from seen classes to unseen classes both difficult and inefficient. (2) Rare labeled novel samples usually cannot provide enough supervision signals to enable the model to adjust from the source distribution to the target distribution, especially for complicated scenarios. To alleviate the above issues, we propose a simple and effective strategy for few-shot and zero-shot text classification. We aim to liberate the model from the confines of seen classes, thereby enabling it to predict unseen categories without the necessity of training on seen classes. Specifically, for mining more related unseen category knowledge, we utilize a large pre-trained language model to generate pseudo novel samples, and select the most representative ones as category anchors. After that, we convert the multi-class classification task into a binary classification task and use the similarities of query-anchor pairs for prediction to fully leverage the limited supervision signals. Extensive experiments on six widely used public datasets show that our proposed method can outperform other strong baselines significantly in few-shot and zero-shot tasks, even without using any seen class samples.

Meibomian glands segmentation in infrared images with limited annotation.

To investigate a pioneering framework for the segmentation of meibomian glands (MGs), using limited annotations to reduce the workload on ophthalmologists and enhance the efficiency of clinical diagnosis. Totally 203 infrared meibomian images from 138 patients with dry eye disease, accompanied by corresponding annotations, were gathered for the study. A rectified scribble-supervised gland segmentation (RSSGS) model, incorporating temporal ensemble prediction, uncertainty estimation, and a transformation equivariance constraint, was introduced to address constraints imposed by limited supervision information inherent in scribble annotations. The viability and efficacy of the proposed model were assessed based on accuracy, intersection over union (IoU), and dice coefficient. Using manual labels as the gold standard, RSSGS demonstrated outcomes with an accuracy of 93.54%, a dice coefficient of 78.02%, and an IoU of 64.18%. Notably, these performance metrics exceed the current weakly supervised state-of-the-art methods by 0.76%, 2.06%, and 2.69%, respectively. Furthermore, despite achieving a substantial 80% reduction in annotation costs, it only lags behind fully annotated methods by 0.72%, 1.51%, and 2.04%. An innovative automatic segmentation model is developed for MGs in infrared eyelid images, using scribble annotation for training. This model maintains an exceptionally high level of segmentation accuracy while substantially reducing training costs. It holds substantial utility for calculating clinical parameters, thereby greatly enhancing the diagnostic efficiency of ophthalmologists in evaluating meibomian gland dysfunction.

EELS: Autonomous snake-like robot with task and motion planning capabilities for ice world exploration.

Ice worlds are at the forefront of astrobiological interest because of the evidence of subsurface oceans. Enceladus in particular is unique among the icy moons because there are known vent systems that are likely connected to a subsurface ocean, through which the ocean water is ejected to space. An existing study has shown that sending small robots into the vents and directly sampling the ocean water is likely possible. To enable such a mission, NASA's Jet Propulsion Laboratory is developing a snake-like robot called Exobiology Extant Life Surveyor (EELS) that can navigate Enceladus' extreme surface and descend an erupting vent to capture unaltered liquid samples and potentially reach the ocean. However, navigating to and through Enceladus' environment is challenging: Because of the limitations of existing orbital reconnaissance, there is substantial uncertainty with respect to its geometry and the physical properties of the surface/vents; communication is limited, which requires highly autonomous robots to execute the mission with limited human supervision. Here, we provide an overview of the EELS project and its development effort to create a risk-aware autonomous robot to navigate these extreme ice terrains/environments. We describe the robot's architecture and the technical challenges to navigate and sense the icy environment safely and effectively. We focus on the challenges related to surface mobility, task and motion planning under uncertainty, and risk quantification. We provide initial results on mobility and risk-aware task and motion planning from field tests and simulated scenarios.

Differential-Critic GAN: Generating What You Want by a Cue of Preferences.

This article proposes differential-critic generative adversarial network (DiCGAN) to learn the distribution of user-desired data when only partial instead of the entire dataset possesses the desired property. DiCGAN generates desired data that meet the user's expectations and can assist in designing biological products with desired properties. Existing approaches select the desired samples first and train regular GANs on the selected samples to derive the user-desired data distribution. However, the selection of the desired data relies on global knowledge and supervision over the entire dataset. DiCGAN introduces a differential critic that learns from pairwise preferences, which are local knowledge and can be defined on a part of training data. The critic is built by defining an additional ranking loss over the Wasserstein GAN's critic. It endows the difference of critic values between each pair of samples with the user preference and guides the generation of the desired data instead of the whole data. For a more efficient solution to ensure data quality, we further reformulate DiCGAN as a constrained optimization problem, based on which we theoretically prove the convergence of our DiCGAN. Extensive experiments on a diverse set of datasets with various applications demonstrate that our DiCGAN achieves state-of-the-art performance in learning the user-desired data distributions, especially in the cases of insufficient desired data and limited supervision.

Integrating coevolutionary strategies and risk preferences: a novel supervision insight for pollutant abatement

The importance of pollutant abatement has been steadily growing in recent times, prompting an increased focus on developing effective regulatory mechanisms. This paper introduces a novel approach by combining theories of evolutionary games and opinion dynamics to formulate a coevolution model of game and preference. Recognizing the challenges posed by limited supervision ability and enterprises’ heterogeneous risk preferences, we propose a smart supervision mechanism. This mechanism incorporates the concepts of whitelist capability and observation period to establish intelligent supervision. Simulation results demonstrate the regulator’s ability to accurately discern enterprises’ preferences based on decision-making differences. The smart supervision mechanism proves to be more effective in achieving pollutant abatement goals compared to random supervision. Furthermore, our findings indicate that with higher supervision ability, increasing whitelist capability enhances cooperation rates. Conversely, lower supervision ability necessitates a shorter observation period and increased whitelist capability to achieve optimal pollutant abatement results. The study highlights that enterprises with a high cooperation rate experience more significant benefits, while risk-seeking enterprises benefit less due to heightened regulator attention at the same cooperation rate.