Explore the vast range of titles available.

News currently spreads rapidly through the internet. Because fake news stories are designed to attract readers, they tend to spread faster. For most readers, detecting fake news can be challenging and such readers usually end up believing that the fake news story is fact. Because fake news can be socially problematic, a model that automatically detects such fake news is required. In this paper, we focus on data-driven automatic fake news detection methods. We first apply the Bidirectional Encoder Representations from Transformers model (BERT) model to detect fake news by analyzing the relationship between the headline and the body text of news. To further improve performance, additional news data are gathered and used to pre-train this model. We determine that the deep-contextualizing nature of BERT is best suited for this task and improves the 0.14 F-score over older state-of-the-art models.

This book describes the theories, applications, and challenges for different oral controlled release formulations.This book differs from most in its focus on oral controlled release formulation design and process development.

Pre-processing and post-processing are significant aspects of natural language processing (NLP) application software. Pre-processing in neural machine translation (NMT) includes subword tokenization to alleviate the problem of unknown words, parallel corpus filtering that only filters data suitable for training, and data augmentation to ensure that the corpus contains sufficient content. Post-processing includes automatic post editing and the application of various strategies during decoding in the translation process. Most recent NLP researches are based on the Pretrain-Finetuning Approach (PFA). However, when small and medium-sized organizations with insufficient hardware attempt to provide NLP services, throughput and memory problems often occur. These difficulties increase when utilizing PFA to process low-resource languages, as PFA requires large amounts of data, and the data for low-resource languages are often insufficient. Utilizing the current research premise that NMT model performance can be enhanced through various pre-processing and post-processing strategies without changing the model, we applied various decoding strategies to Korean–English NMT, which relies on a low-resource language pair. Through comparative experiments, we proved that translation performance could be enhanced without changes to the model. We experimentally examined how performance changed in response to beam size changes and n-gram blocking, and whether performance was enhanced when a length penalty was applied. The results showed that various decoding strategies enhance the performance and compare well with previous Korean–English NMT approaches. Therefore, the proposed methodology can improve the performance of NMT models, without the use of PFA; this presents a new perspective for improving machine translation performance.

It is generally assumed that polymeric micelles, upon administration into the blood stream, carry drug molecules until they are taken up into cells followed by intracellular release. The current work revisits this conventional wisdom. The study using dual-labeled micelles containing fluorescently labeled copolymers and hydrophobic fluorescent probes entrapped in the polymeric micelle core showed that cellular uptake of hydrophobic probes was much faster than that of labeled copolymers. This result implies that the hydrophobic probes in the core are released from micelles in the extracellular space. Förster resonance energy transfer (FRET) imaging and spectroscopy were used to monitor this process in real time. A FRET pair, DiIC₁₈₍₃₎ and DiOC₁₈₍₃₎, was loaded into monomethoxy poly(ethylene glycol)-block-poly(D,L-lactic acid) micelles. By monitoring the FRET efficiency, release of the core-loaded probes to model membranes was demonstrated. During administration of polymeric micelles to tumor cells, a decrease of FRET was observed both on the cell membrane and inside of cells, indicating the release of core-loaded probes to the cell membrane before internalization. The decrease of FRET on the plasma membrane was also observed during administration of paclitaxel-loaded micelles. Taken together, our results suggest a membrane-mediated pathway for cellular uptake of hydrophobic molecules preloaded in polymeric micelles. The plasma membrane provides a temporal residence for micelle-released hydrophobic molecules before their delivery to target intracellular destinations. A putative role of the PEG shell in the molecular transport from micelle to membrane is discussed.

The differential immune responses after two additional BNT162b2 (BNT) booster doses between ChAdOx1 nCoV-10 (ChAd)-primed and BNT-primed groups have not been elucidated. The aim of this study was to compare vaccine-induced humoral and cellular immune responses and evaluate breakthrough infection between the two vaccination strategies. In 221 healthy subjects (111 in the ChAd group), longitudinal immune responses were monitored at 3, 4, and 6 months after the 2nd dose and 1, 3, and 6 months after the 3rd dose. Humoral immunity was measured by two fully automated chemiluminescent immunoassays (Elecsys and Abbott) and a surrogate virus neutralization test (sVNT). Cellular immunity was assessed by two interferon-γ (IFN-γ) release assays (QuantiFERON SARS-CoV-2 and Covi-FERON). After the 2nd dose of BNT vaccination, total antibody levels were higher in the ChAd group, but IgG antibody and sVNT results were higher in the BNT group. Following the 3rd dose vaccination, binding antibody titers were significantly elevated in both groups (ChAD-BNT; 15.4 to 17.8-fold, BNT-BNT; 22.2 to 24.6-fold), and the neutralizing capacity was increased by 1.3-fold in both cohorts. The ChAd-BNT group had lower omicron neutralization positivity than the BNT-BNT group (P = 0.001) at 6 months after the 3rd dose. Cellular responses to the spike antigen also showed 1.7 to 3.0-fold increases after the 3rd dose, which gradually declined to the levels equivalent to before the 3rd vaccination. The ChAd cohort tended to have higher IFN-γ level than the BNT cohort for 3-6 months after the 2nd and 3rd doses. The frequency of breakthrough infection was higher in the ChAd group (44.8%) than in the BNT group (28.1%) (P = 0.0219). Breakthrough infection induced increased humoral responses in both groups, and increase of cellular response was significant in the ChAd group. Our study showed differential humoral and cellular immune responses between ChAd-BNT-BNT heterologous and BNT-BNT-BNT homologous vaccination cohorts. The occurrence of low antibody levels in the ChAd-primed cohort in the humoral immune response may be associated with an increased incidence of breakthrough infections. Further studies are needed on the benefits of enhanced cellular immunity in ChAd-primed cohorts.

Silymarin has been widely used as a hepatoprotective drug in the treatment of various liver diseases, yet its effectiveness is affected by its poor water solubility and low bioavailability after oral administration, and there is a need for the development of intravenous products, especially for liver-targeting purposes. In this study, silymarin was encapsulated in self-assembled nanoparticles of Bletilla striata polysaccharide (BSP) conjugates modified with stearic acid and the physicochemical properties of the obtained nanoparticles were characterized. The silymarin-loaded micelles appeared as spherical particles with a mean diameter of 200 nm under TEM. The encapsulation of drug molecules was confirmed by DSC thermograms and XRD diffractograms, respectively. The nanoparticles exhibited a sustained-release profile for nearly 1 week with no obvious initial burst. Compared to drug solutions, the drug-loaded nanoparticles showed a lower viability and higher uptake intensity on HepG2 cell lines. After intravenous administration of nanoparticle formulation for 30 min to mice, the liver became the most significant organ enriched with the fluorescent probe. These results suggest that BSP derivative nanoparticles possess hepatic targeting capability and are promising nanocarriers for delivering silymarin to the liver.

The commonsense question and answering (CSQA) system predicts the right answer based on a comprehensive understanding of the question. Previous research has developed models that use QA pairs, the corresponding evidence, or the knowledge graph as an input. Each method executes QA tasks with representations of pre-trained language models. However, the ability of the pre-trained language model to comprehend completely remains debatable. In this study, adversarial attack experiments were conducted on question-understanding. We examined the restrictions on the question-reasoning process of the pre-trained language model, and then demonstrated the need for models to use the logical structure of abstract meaning representations (AMRs). Additionally, the experimental results demonstrated that the method performed best when the AMR graph was extended with ConceptNet. With this extension, our proposed method outperformed the baseline in diverse commonsense-reasoning QA tasks.

BlenderBot 2.0 represents a significant advancement in open-domain chatbots by incorporating real-time information and retaining user information across multiple sessions through an internet search module. Despite its innovations, there are still areas for improvement. This paper examines BlenderBot 2.0’s limitations and errors from three perspectives: model, data, and user interaction. From the data perspective, we highlight the challenges associated with the crowdsourcing process, including unclear guidelines for workers, insufficient measures for filtering hate speech, and the lack of a robust process for verifying the accuracy of internet-sourced information. From the user perspective, we identify nine types of limitations and conduct a thorough investigation into their causes. For each perspective, we propose practical methods for improvement and discuss potential directions for future research. Additionally, we extend our analysis to include perspectives in the era of large language models (LLMs), further broadening our understanding of the challenges and opportunities present in current AI technologies. This multifaceted analysis not only sheds light on BlenderBot 2.0’s current limitations but also charts a path forward for the development of more sophisticated and reliable open-domain chatbots within the broader context of LLM advancements.

Unlike previous dialogue-based question-answering (QA) datasets, DREAM, multiple-choice Dialogue-based REAding comprehension exaMination dataset, requires a deep understanding of dialogue. Many problems require multi-sentence reasoning, whereas some require commonsense reasoning. However, most pre-trained language models (PTLMs) do not consider commonsense. In addition, because the maximum number of tokens that a language model (LM) can deal with is limited, the entire dialogue history cannot be included. The resulting information loss has an adverse effect on performance. To address these problems, we propose a Dialogue-based QA model with Common-sense Reasoning (DQACR), a language model that exploits Semantic Search and continual learning. We used Semantic Search to complement information loss from truncated dialogue. In addition, we used Semantic Search and continual learning to improve the PTLM’s commonsense reasoning. Our model achieves an improvement of approximately 1.5% over the baseline method and can thus facilitate QA-related tasks. It contributes toward not only dialogue-based QA tasks but also another form of QA datasets for future tasks.

Cervical spondylosis is a common cause of spinal cord dysfunction, and anterior cervical discectomy and fusion (ACDF) is widely employed when conservative treatment fails. Conventional implant systems such as the cervical cage with plate (CCP) and zero-profile stand-alone cage (ZPSC) are commonly used to enhance spinal stability and promote fusion, but they are associated with complications including dysphagia and adjacent segment degeneration. To address these limitations, a novel flexible plate cage system (FPCS) has been developed to optimize biomechanical performance while minimizing surgical risk. In this study, a finite element model of the C3–T1 cervical spine was constructed to simulate ACDF at the C5–C6 level using CCP, ZPSC, and FPCS implants. Under standardized loading conditions, von Mises stress was analyzed in the bone, intervertebral disc, endplates, cage, and screws, using the mean of the top 5% stress values to ensure accuracy. All surgical models showed increased stress compared to the intact reference spine. The ZPSC model exhibited the highest stress in the cage and screws, suggesting a more concentrated load path. The CCP model showed a more evenly distributed stress profile, particularly affecting the inferior adjacent segment. The FPCS model demonstrated moderate cage stress, reduced screw stress, and the highest plate stress, indicating a design that effectively redirects mechanical load from the screw-bone interface toward the anterior plate. This may be related to the unique structural configuration of the FPCS, which secures screws horizontally into the anterior vertebral body without penetrating the endplates. These findings suggest that the FPCS may offer a biomechanically favorable alternative to existing ACDF implants.