Specialised pre-trained language models are becoming more frequent in Natural language Processing (NLP) since they can potentially outperform models trained on generic texts. BioBERT (Sanh et al., Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. arXiv preprint arXiv: 1910.01108, 2019) and BioClinicalBERT (Alsentzer et al., Publicly available clinical bert embeddings. In Proceedings of the 2nd Clinical Natural Language Processing Workshop, pp. 72–78, 2019) are two examples of such models that have shown promise in medical NLP tasks. Many of these models are overparametrised and resource-intensive, but thanks to techniques like knowledge distillation, it is possible to create smaller versions that perform almost as well as their larger counterparts. In this work, we specifically focus on development of compact language models for processing clinical texts (i.e. progress notes, discharge summaries, etc). We developed a number of efficient lightweight clinical transformers using knowledge distillation and continual learning, with the number of parameters ranging from $15$ million to $65$ million. These models performed comparably to larger models such as BioBERT and ClinicalBioBERT and significantly outperformed other compact models trained on general or biomedical data. Our extensive evaluation was done across several standard datasets and covered a wide range of clinical text-mining tasks, including natural language inference, relation extraction, named entity recognition and sequence classification. To our knowledge, this is the first comprehensive study specifically focused on creating efficient and compact transformers for clinical NLP tasks. The models and code used in this study can be found on our Huggingface profile at https://huggingface.co/nlpie and Github page at https://github.com/nlpie-research/Lightweight-Clinical-Transformers, respectively, promoting reproducibility of our results.

Motion planning for high-DOF multi-arm systems operating in complex environments remains a challenging problem, with many motion planning algorithms requiring evaluation of the minimum collision distance and its derivative. Because of the computational complexity of calculating the collision distance, recent methods have attempted to leverage data-driven machine learning methods to learn the collision distance. Because of the significant training dataset requirements for high-DOF robots, existing kernel-based methods, which require $O(N^2)$ memory and computation resources, where $N$ denotes the number of dataset points, often perform poorly. This paper proposes a new active learning method for learning the collision distance function that overcomes the limitations of existing methods: (i) the size of the training dataset remains fixed, with the dataset containing more points near the collision boundary as learning proceeds, and (ii) calculating collision distances in the higher-dimensional link $SE(3)^n$ configuration space – here $n$ denotes the number of links – leads to more accurate and robust collision distance function learning. Performance evaluations with high-DOF multi-arm robot systems demonstrate the advantages of the proposed active learning-based strategy vis-$\grave{\text{a}}$-vis existing learning-based methods.

Viral marketing campaigns target primarily those individuals who are central in social networks and hence have social influence. Marketing events, however, may attract diverse audience. Despite the importance of event marketing, the influence of heterogeneous target groups is not well understood yet. In this paper, we define the Audience Selection (AS) problem in which different sets of agents need to be evaluated and compared based on their social influence. A typical application of Audience selection is choosing locations for a series of marketing events. The Audience selection problem is different from the well-known Influence Maximization (IM) problem in two aspects. Firstly, it deals with sets rather than nodes. Secondly, the sets are diverse, composed by a mixture of influential and ordinary agents. Thus, Audience selection needs to assess the contribution of ordinary agents too, while IM only aims to find top spreaders. We provide a systemic test for ranking influence measures in the Audience Selection problem based on node sampling and on a novel statistical method, the Sum of Ranking Differences. Using a Linear Threshold diffusion model on two online social networks, we evaluate eight network measures of social influence. We demonstrate that the statistical assessment of these influence measures is remarkably different in the Audience Selection problem, when low-ranked individuals are present, from the IM problem, when we focus on the algorithm’s top choices exclusively.

This article considers the individual equilibrium behavior and socially optimal strategy in a fluid queue with two types of parallel customers and incomplete fault. Assume that the working state and the incomplete fault state appear alternately in the buffer. Different from the linear revenue and expenditure structure, an exponential utility function can be constructed to obtain the equilibrium balking thresholds in the fully observable case. Besides, the steady-state probability distribution and the corresponding expected social benefit are derived based on the renewal process and the standard theory of linear ordinary differential equations. Furthermore, a reasonable entrance fee strategy is discussed under the condition that the fluid accepts the globally optimal strategies. Finally, the effects of the diverse system parameters on the entrance fee and the expected social benefit are explicitly illustrated by numerical comparisons.

We study 2-stage game-theoretic problem oriented 3-stage service policy computing, convolutional neural network (CNN) based algorithm design, and simulation for a blockchained buffering system with federated learning. More precisely, based on the game-theoretic problem consisting of both “win-lose” and “win-win” 2-stage competitions, we derive a 3-stage dynamical service policy via a saddle point to a zero-sum game problem and a Nash equilibrium point to a non-zero-sum game problem. This policy is concerning users-selection, dynamic pricing, and online rate resource allocation via stable digital currency for the system. The main focus is on the design and analysis of the joint 3-stage service policy for given queue/environment state dependent pricing and utility functions. The asymptotic optimality and fairness of this dynamic service policy is justified by diffusion modeling with approximation theory. A general CNN based policy computing algorithm flow chart along the line of the so-called big model framework is presented. Simulation case studies are conducted for the system with three users, where only two of the three users can be selected into the service by a zero-sum dual cost game competition policy at a time point. Then, the selected two users get into service and share the system rate service resource through a non-zero-sum dual cost game competition policy. Applications of our policy in the future blockchain based Internet (e.g., metaverse and web3.0) and supply chain finance are also briefly illustrated.

This paper proposes a robust control approach to achieve high-precision trajectory tracking for permanent magnet linear motor (PMLM) system containing uncertainties by describing the dynamic model of PMLM based on the Udwadia-Kalaba equation combined with constraint-following method. First, the system of PMLM is described as a constraint-following system by adding the generalized constraint force to the unconstrained Udwadia-Kalaba equation of PMLM system. Second, the robust constraint-following controller is designed based on the proposed model after uncertainty analysis. Moreover, the proposed controller is verified to guarantee deterministic performance for uncertain systems: uniformly bounded and uniformly ultimately bounded. Third, the numerical simulation and experimental validation demonstrate the effectiveness of proposed controller. Finally, the design approach of constraint-following can be applied to other systems with uncertainties.

Simultaneous localization and mapping systems based on rigid scene assumptions cannot achieve reliable positioning and mapping in a complex environment with many moving objects. To solve this problem, this paper proposes a novel dynamic multi-object lidar odometry (MLO) system based on semantic object recognition technology. The proposed system enables the reliable localization of robots and semantic objects and the generation of long-term static maps in complex dynamic scenes. For ego-motion estimation, the proposed system extracts environmental features that take into account both semantic and geometric consistency constraints. Then, the filtered features can be robust to the semantic movable and unknown dynamic objects. In addition, we propose a new least-squares estimator that uses geometric object points and semantic box planes to realize the multi-object tracking (SGF-MOT) task robustly and precisely. In the mapping module, we implement dynamic semantic object detection using the absolute trajectory tracking list. By using static semantic objects and environmental features, the system eliminates accumulated localization errors and produces a purely static map. Experiments on the public KITTI dataset show that the proposed MLO system provides more accurate and robust object tracking performance and better real-time localization accuracy in complex scenes compared to existing technologies.

Data on real-time individuals’ location may provide significant opportunities for managing emergency situations. For example, in the case of outbreaks, besides informing on the proximity of people, hence supporting contact tracing activities, location data can be used to understand spatial heterogeneity in virus transmission. However, individuals’ low consent to share their data, proved by the low penetration rate of contact tracing apps in several countries during the coronavirus disease-2019 (COVID-19) pandemic, re-opened the scientific and practitioners’ discussion on factors and conditions triggering citizens to share their positioning data. Following the Antecedents → Privacy Concerns → Outcomes (APCO) model, and based on Privacy Calculus and Reasoned Action Theories, the study investigates factors that cause university students to share their location data with public institutions during outbreaks. To this end, an explanatory survey was conducted in Italy during the second wave of COVID-19, collecting 245 questionnaire responses. Structural equations modeling was used to contemporary investigate the role of trust, perceived benefit, and perceived risk as determinants of the intention to share location data during outbreaks. Results show that respondents’ trust in public institutions, the perceived benefits, and the perceived risk are significant predictor of the intention to disclose personal tracking data with public institutions. Results indicate that the latter two factors impact university students’ willingness to share data more than trust, prompting public institutions to rethink how they launch and manage the adoption process for these technological applications.

Bus Rapid Transit (BRT) has grown fast in the last 25 years, promising low-cost, rapid implementation, and large positive impacts. Despite advances, many systems in middle- and low-income countries face operational and financial issues, particularly in Latin America. Some practitioners, researchers, and decision makers, and the media are questioning its ability to provide quality services. Is this the end of a trend? To answer this question, this paper explores the status of the BRT industry and literature on the topic, with a focus on Latin America, as well as the emblematic cases of Curitiba, Quito, Bogotá, Mexico, and Santiago. Overcrowding, lack of reliability, fare evasion, issues of safety and security, and poor maintenance are evident problems in these and other cities. They seem to be a result of institutional and financial constraints, as well as technical limitations of surface-based transit modes. BRT has been able to deliver high-capacity and fast and reliable services, but requires permanent management and investment to face growing demand and aging infrastructure and vehicles, just like rail systems do. In addition, attention needs to be provided to data, technology innovation, urban integration, and public participation to keep BRT as an integral part of multimodal high-quality sustainable mobility networks in the future.

In this paper, we introduce a novel way to quantify the remaining inaccuracy of order statistics by utilizing the concept of extropy. We explore various properties and characteristics of this new measure. Additionally, we expand the notion of inaccuracy for ordered random variables to a dynamic version and demonstrate that this dynamic information measure provides a unique determination of the distribution function. Moreover, we investigate specific lifetime distributions by analyzing the residual inaccuracy of the first-order statistics. Nonparametric kernel estimation of the proposed measure is suggested. Simulation results show that the kernel estimator with bandwidth selection using the cross-validation method has the best performance. Finally, an application of the proposed measure on the model selection is provided.

We give a construction of the free dcpo-cone over any dcpo. There are two steps for getting this result. Firstly, we extend the notion of power domain to directed spaces which are equivalent to $T_0$ monotone-determined spaces introduced by Erné, and we construct the probabilistic powerspace of the monotone determined space, which is defined as a free monotone determined cone. Secondly, we take D-completion of the free monotone determined cone over the dcpo with its Scott topology. In addition, we show that generally the valuation power domain of any dcpo is not the free dcpo-cone.