In this paper, we investigate the modeling power of contextualized embeddings from pretrained language models, e.g. BERT, on the E2E-ABSA task. Specifically, we build a series of simple yet insightful neural baselines...
详细信息
We review simulation optimization methods and their connection with artificial intelligence (AI) techniques. In particular, we focus on two areas: stochastic gradient estimation, which plays a central role in training...
详细信息
We review simulation optimization methods and their connection with artificial intelligence (AI) techniques. In particular, we focus on two areas: stochastic gradient estimation, which plays a central role in training neural networks for deep learning and reinforcement learning; and ranking and selection, which can be used as the node selection policy in Monte Carlo tree search. We also review the literature on inventory management, which has been studied in both simulation optimization and AI.
Debates continue over the effectiveness of limiting alcohol outlet density in reducing alcohol consumption, and its broader impacts on access to non-alcoholic services in low-income urban communities remain underexplo...
详细信息
The use of Social Network Centrality alongside Content Centric Networking is used towards improving user's access to content, as well as maximizing the efficient exploitation of network resources. This work demons...
详细信息
This research explores anthropomorphism and gender presentation as prospective determinants of trust in household service robots with respect to care of objects (e.g., clothing, valuables), information (e.g., online p...
详细信息
This research explores anthropomorphism and gender presentation as prospective determinants of trust in household service robots with respect to care of objects (e.g., clothing, valuables), information (e.g., online passwords, credit card numbers), and living agents (e.g., pets, children). In Experiments 1 and 2, we compared trust in a humanoid robot presenting as male, female, or gender-neutral, finding no effects of gender presentation on any trust outcome. In Experiment 3, a fourth condition depicting a physically nonhumanoid robot was added. Relative to the humanoid conditions, participants reported less willingness to trust the nonhumanoid robot to care for their objects, personal information, or vulnerable agents; the reduced trust in care for objects or information was mediated by appraisals of the nonhumanoid as less intelligent and less likable, whereas the reduced trust in care of agents was mediated by appraisals of the nonhumanoid as less likable and less alive. In a parallel pattern, across all studies, participants’ appraisals of robots as intelligent tracked trust in them to take care of objects or information (but not agents), whereas appraisals of robots as likable and alive tracked trust in care of agents. The results are discussed as they inform past work examining effects of gender presentation and anthropomorphism on perceptions of, and trust in, robots.
Bangladesh ranks among the top 10 countries globally in terms of climate change impacts and faces numerous anthropogenic and natural pressures. Cox's Bazar, its primary tourist district, is experiencing severe deg...
详细信息
Bangladesh ranks among the top 10 countries globally in terms of climate change impacts and faces numerous anthropogenic and natural pressures. Cox's Bazar, its primary tourist district, is experiencing severe degradation of its physical and ecological environments due to anthropogenic disturbances and climate change. To improve its environmental quality and preserve its ecological resources effectively, it is essential to develop a spatial decision support instrument addressing multi-pressures and cumulative environmental vulnerability (EV). This study presents an expert opinion-independent, scalable, and customizable spatial methodological framework, integrating multi-sourced geospatial data with GIS-based Fuzzy Logic to assess spatial distributions of five pressure groups and their resulting EV in Cox's Bazar. 18 criteria were chosen based on a structured literature review to evaluate the five pressure groups. Results revealed that 17 % to 27 % of the study area is exposed to high to very high hydro-meteorological, topographic, land resource, anthropogenic, and natural hazard pressures. The EV results indicated that one-third of the study area, majorly covering Kutubdia, Pekua, Cox's Bazar Sadar, Teknaf, and Ukhia upazilas, is highly environmentally vulnerable. For enhanced environmental protection, this study improved the existing method of environmental protection zoning by introducing a novel zoning approach that integrates in-situ biodiversity data with EV data. This new zoning method delineated 24 % (555 sq. km.) as strict, 45 % (1047 sq. km.) as medium, and 31 % (725 sq. km.) as soft protection zones in the study area. The sensitivity analysis identified land resource pressure as the most influential component of EV. With a correlation coefficient of 0.91, the accuracy assessment confirms a high level of reliability in the EV results. This study provides valuable insights into environmental pressures and vulnerability in Cox's Bazar, which are crucial for i
Offshore wind power (OWP) has become one of the most important renewable energy sources in the global energy transition. The interactive influence of different policies, i.e., Feed-in tariff (FIT), R&D subsidies, ...
详细信息
The cell nuclei segmentation of is a challenging task in microscopy image analysis. The problems of noise, small cell nuclei, and few training data samples in the data set will all affect the effectiveness of the mode...
详细信息
Motivated by modern regression applications, in this paper, we study the convexification of a class of convex optimization problems with indicator variables and combinatorial constraints on the indicators. Unlike most...
详细信息
We study a risk-sharing economy where an arbitrary number of heterogenous agents trades an arbitrary number of risky assets subject to quadratic transaction costs. For linear state dynamics, the forward-backward stoch...
详细信息
暂无评论