RefAI: a GPT-powered retrieval-augmented generative tool for biomedical literature recommendation and summarization.

OBJECTIVES: Precise literature recommendation and summarization are crucial for biomedical professionals. While the latest iteration of generative pretrained transformer (GPT) incorporates 2 distinct modes-real-time search and pretrained model utilization-it encounters challenges in dealing with these tasks. Specifically, the real-time search can pinpoint some relevant articles but occasionally provides fabricated papers, whereas the pretrained model excels in generating well-structured summaries but struggles to cite specific sources. In response, this study introduces RefAI, an innovative retrieval-augmented generative tool designed to synergize the strengths of large language models (LLMs) while overcoming their limitations. MATERIALS AND METHODS: RefAI utilized PubMed for systematic literature retrieval, employed a novel multivariable algorithm for article recommendation, and leveraged GPT-4 turbo for summarization. Ten queries under 2 prevalent topics ("cancer immunotherapy and target therapy" and "LLMs in medicine") were chosen as use cases and 3 established counterparts (ChatGPT-4, ScholarAI, and Gemini) as our baselines. The evaluation was conducted by 10 domain experts through standard statistical analyses for performance comparison. RESULTS: The overall performance of RefAI surpassed that of the baselines across 5 evaluated dimensions-relevance and quality for literature recommendation, accuracy, comprehensiveness, and reference integration for summarization, with the majority exhibiting statistically significant improvements (P-values <.05). DISCUSSION: RefAI demonstrated substantial improvements in literature recommendation and summarization over existing tools, addressing issues like fabricated papers, metadata inaccuracies, restricted recommendations, and poor reference integration. CONCLUSION: By augmenting LLM with external resources and a novel ranking algorithm, RefAI is uniquely capable of recommending high-quality literature and generating well-structured summaries, holding the potential to meet the critical needs of biomedical professionals in navigating and synthesizing vast amounts of scientific literature.

Single-metal atom sandwiched by graphdiyne and BN-doped graphdiyne sheets as an electrocatalyst for nitrogen reduction: a first-principles study.

The exploration of efficient single-atom catalysts provides a prospective pattern for the sustainable development of electrocatalytic nitrogen fixation. We systematically researched the nitrogen reduction properties of catalysts with a single transition metal (TM) atom sandwiched between BN-doped graphdiyne and graphdiyne (labeled BN-TM-G) by first-principles calculations. The TM atom in the novel sandwich structure provides electrons to the adjacent B atom, which acts as the active site, thus driving the fixation and reduction of N2. In the BN-TM-G system, the NRR catalytic activity is bound up with the positive charge polarization level of the TM atom. Among them, BN-Sc-G, BN-Ti-G, BN-V-G, and BN-Cr-G systems showed higher catalytic ability, and the competitive HER was inhibited. In particular, the lowest limiting potential of BN-Cr-G is -0.63 V is promising for the NRR catalyst.

Neurological complications of COVID-19 in children and the associated immunological responses.

The high spread rate, severe symptoms, psychological and neurological problems, and unavailability of effective medicines are the major factors making Coronavirus disease 2019 (COVID-19) a massive threat to the world. It is thought that COVID-19 causes mild symptoms or mild infectious illness in children. However, we cannot rule out the possibility of serious complications such as the multisystem inflammatory syndrome. COVID-19 induces mild to severe neurological problems in children, such as stroke, encephalopathy, mild shortness of breath, and myalgia. The development of these conditions is associated with pro-inflammatory responses and cytokine storms, which alter the physiology of the blood-brain barrier and allow the virus to enter the brain. Despite the viral entry into the brain, these neurological conditions can also be caused indirectly by severe immune responses. In this article, we describe COVID-19 and the associated neurological and immunological complications in children.

Multiscale dual attention mechanism for fluid segmentation of optical coherence tomography images.

Optical coherence tomography (OCT) technology can obtain a clear retinal structure map, which is greatly beneficial for the diagnosis of retinopathy. Ophthalmologists can use OCT technology to analyze information about the retina's internal structure and changes in retinal thickness. Therefore, segmentation of retinal layers in images and screening for retinal diseases have become important goals in OCT scanning. In this paper, we propose the multiscale dual attention (MSDA)-UNet network, an MSDA mechanism network for OCT lesion area segmentation. The MSDA-UNet network introduces position and multiscale channel attention modules to calculate a global reference for each pixel prediction. The network can extract the lesion area information of OCT images of different scales and perform end-to-end segmentation of the OCT retinopathy area. The network framework was trained and tested on the same OCT dataset and compared with other OCT fluid segmentation methods to assess its effectiveness.

Comparison of Telemedicine Versus In-Person Visits on Impact of Downstream Utilization of Care.

Background: Telemedicine use has expanded substantially in recent years. Studies evaluating the impact of telemedicine modalities on downstream office visits have demonstrated mixed results. Introduction: We evaluated insurance claims of a large commercial payer, Blue Cross Blue Shield of Michigan (BCBSM), to assess the frequency of follow-up visits following encounters initiated via telemedicine versus in-person. Materials and Methods: We used the BCBSM claim-level data set (2011-2017) to assess encounters in the following places of service: hospital outpatient, doctor's office, patient's home, or psychiatric daycare facility. We identified the primary diagnostic category for 30-day episodes of care using clinical classifications software (CCS) and multilevel clinical classifications software (ML-CCS). Our intervention group consisted of episodes initiated via telemedicine; our control group consisted of episodes initiated in-person. Our primary outcome was the percentage of 30-day episodes with a related visit (encounters occurring within the same period and CCS categories) across CCS categories. Our secondary outcome was the mean related visit rate. Results: The final data set included 4,982,456 patients and 68,148,070 claims, of which 53,853 were telemedicine related. Many episodes did not have related visits (the mean related visit rate was 16%). Telemedicine visits had a higher frequency of related visits across all CCS categories. Discussion: Episodes of care initiated via telemedicine more frequently generate related visits within a 30-day period. This increased health care utilization could represent excessive care or could reflect expanded access to care. Conclusion: Further research should explore the cause of this increased utilization and potential unintended consequences.

How urban agglomeration improve the emission efficiency？A spatial econometric analysis of the Yangtze River Delta urban agglomeration in China.

Urban areas consume more than 66% of the world's energy and generate more than 70% of global greenhouse gas (GHG) emissions. With the world's population expected to reach 10 billion by 2100, and with nearly 90% of people living in urban areas, a critical question for planetary sustainability is how the size of cities affects energy use and carbon dioxide (CO2) emissions. Are urban agglomerations more energy and emission efficient than smaller cities? Does urban agglomeration exhibit gains from economies of scale concerning emissions? Here, we examine the relationship between urban agglomeration and CO2 emissions for urban agglomeration in the Yangtze River Delta in China using a STIRPAT (Stochastic Impacts by Regression on Population, Affluence and Technology) model considering the spatial effects. Also, it examines the influence of economic development, industrial structure, opening-up level, and technology progress on carbon emissions by exploring the spatial agglomeration and spillover effects. Our major finding is that urban size has a negative correlation to carbon emissions, demonstrating that urban agglomeration is more emission efficient. In addition, our results showed that carbon emission driving factors, such as technology progress, opening-up, and population, have spatial dependence and spatial spillover effects. It means a city's carbon emissions are not only influenced by its own factors but also have an impact on neighboring cities. Therefore, cross-city or urban agglomeration policy, and actions of reducing carbon emissions, are necessary, whilst also developing a low-carbon economy by increasing the proportion of high-tech industry through technological progress and developing vigorous resource-saving and an environmentally friendly tertiary industry.