– Antennas fed by waveguides with irregular cross-sections are efficiently simulated through the hybridization of the mode-matching (MM) method and the higher-order method of moments (HOMoM). To obtain the waveguide ...
详细信息
The increaing significance of plant life and botanical expertise extends beyond mere visual appreciation. With the growing interest in sustainable living and alternative remedies, there is a pressing demand for easily...
详细信息
In today’s rapidly changing world, cloud service providers face numerous challenges in managing resources and meeting customer demands. To address these challenges, cloud service providers should prioritize the tasks...
详细信息
In the realm of clinical healthcare, medical visual question answering systems emerge as a pivotal innovation that plays a crucial role in clinical decision-making and patient care. They are designed to interpret medi...
详细信息
Test-time adaptation (TTA) seeks to tackle potential distribution shifts between training and testing data by adapting a given model w.r.t. any testing sample. This task is particularly important when the test environ...
详细信息
Test-time adaptation (TTA) seeks to tackle potential distribution shifts between training and testing data by adapting a given model w.r.t. any testing sample. This task is particularly important when the test environment changes frequently. Although some recent attempts have been made to handle this task, we still face two key challenges: 1) prior methods have to perform backpropagation for each test sample, resulting in unbearable optimization costs to many applications;2) while existing TTA solutions can significantly improve the test performance on out-of-distribution data, they often suffer from severe performance degradation on in-distribution data after TTA (known as catastrophic forgetting). To this end, we have proposed an Efficient Anti-Forgetting Test-Time Adaptation (EATA) method which develops an active sample selection criterion to identify reliable and non-redundant samples for test-time entropy minimization. To alleviate forgetting, EATA introduces a Fisher regularizer estimated from test samples to constrain important model parameters from drastic changes. However, in EATA, the adopted entropy loss consistently assigns higher confidence to predictions even when the samples are underlying uncertain, leading to overconfident predictions that underestimate the data uncertainty. To tackle this, we further propose EATA with Calibration (EATA-C) to separately exploit the reducible model uncertainty and the inherent data uncertainty for calibrated TTA. Specifically, we compare the divergence between predictions from the full network and its sub-networks to measure the reducible model uncertainty, on which we propose a test-time uncertainty reduction strategy with divergence minimization loss to encourage consistent predictions instead of overconfident ones. To further re-calibrate predicting confidence on different samples, we utilize the disagreement among predicted labels as an indicator of the data uncertainty. Based on this, we devise a min-max entropy
This study presents a framework for evaluating the performance of MEA-1DMA2P solution in CO2 absorption for a structured packed bed by integrating a rate-based model with statistical optimization. The developed model ...
详细信息
Early identification of skin cancer is mandatory to minimize the worldwide death rate as this disease is covering more than 30% of mortality rates in young and adults. Researchers are in the move of proposing advanced...
详细信息
This study presents a comprehensive approach to developing a domain-specific large language model (LLM) for regulatory and financial text interpretation. A specialized corpus was constructed through large-scale scrapi...
详细信息
One of the main challenges facing businesses migrating to the cloud is getting an estimate of their costs in advance. The estimators available to date allow companies to compare the different virtual machine offerings...
详细信息
Sentiment analysis is a subset of NLP and has encountered an outstanding change with the introduction of new approaches to increase its precision and performance. Neural networks and transformers in deep learning are ...
详细信息
暂无评论