Generalized feed-forward Gaussian models have achieved significant progress in sparse-view 3D reconstruction by leveraging prior knowledge from large multi-view datasets. However, these models often struggle to represent high-frequency details due to the limited number of Gaussians. While the densification strategy used in per-scene 3D Gaussian splatting (3D-GS) optimization can be adapted to the feed-forward models, it may not be ideally suited for generalized scenarios. In this paper, we propose Generative Densification, an efficient and generalizable method to densify Gaussians generated by feed-forward models. Unlike the 3D-GS densification strategy, which iteratively splits and clones raw Gaussian parameters, our method up-samples feature representations from the feed-forward models and generates their corresponding fine Gaussians in a single forward pass, leveraging the embedded prior knowledge for enhanced generalization. Experimental results on both object-level and scene-level reconstruction tasks demonstrate that our method outperforms state-of-the-art approaches with comparable or smaller model sizes, achieving notable improvements in representing fine details.
Viruses play a regulatory role in microbial ecology. Traditional fermented foods have complex fermentation environments with abundant viral participation, yet current research on viral communities in fermented foods remains insufficient. Traditional, manually produced solid-state fermented vinegar serves as an excellent model for studying the role of viral communities in fermented foods. Using metagenomic approaches, this study investigates the structure and dynamics of viral communities during the acetic acid fermentation process of Shanxi aged vinegar. All identified viruses were bacteriophages, and the dominant families were identified as Herelleviridae, Autographiviridae, and Stanwilliamsviridae. The richness and diversity of viral communities exhibited significant variations during acetic acid fermentation. Furthermore, correlation analysis revealed a strong association (p < 0.01) between core bacteria and core viruses. Functional annotation revealed the presence of viral genes associated with amino acid and carbohydrate metabolism. Notably, abundant auxiliary carbohydrate-active enzyme (CAZyme) genes were identified in viruses, with glycoside hydrolases (GHs), glycosyltransferases (GTs), and carbohydrate-binding modules (CBMs) demonstrating particularly high abundance. Additionally, several antibiotic resistance genes were detected in viruses. This study elucidates the impact of viral communities on microbial dynamics during food fermentation, advancing our understanding of viral roles in traditional fermented food ecosystems.
Wine is a high-value alcoholic beverage welcomed by consumers because of its flavor and nutritional value. The key information on wine bottle label is the basis of consumers' choice, which also becomes a target for manufacturers to adulterate, including geographical origin, grape variety and vintage. With the improvement of wine adulteration technology, modern technological means are needed to solve the above mentioned problems. The chemical basis of wine determines the type of technique used. Detection technology can be subdivided into four groups: mass spectrometry techniques, spectroscopic techniques, chromatography techniques, and other techniques. Multivariate statistical analysis of the data was performed by means of chemometrics methods. This paper outlines a series of procedures for wine classification and identification, and classified the analytical techniques and data processing methods used in recent years with listing their principles, advantages and disadvantages to help wine researchers choose appropriate methods to meet the challenge and ensure wine traceability and authenticity.
Machine learning (ML) methods have experienced significant growth in the past decade, yet their practical application in high-impact real-world domains has been hindered by their opacity. When ML methods are responsible for making critical decisions, stakeholders often require insights into how to alter these decisions. Counterfactual explanations (CFEs) have emerged as a solution, offering interpretations of opaque ML models and providing a pathway to transition from one decision to another. However, most existing CFE methods require access to the model's training dataset, few methods can handle multivariate time-series, and none can handle multivariate time-series without training datasets. These limitations can be formidable in many scenarios. In this paper, we present CFWoT, a novel reinforcement-learning-based CFE method that generates CFEs when training datasets are unavailable. CFWoT is model-agnostic and suitable for both static and multivariate time-series datasets with continuous and discrete features. Users have the flexibility to specify non-actionable, immutable, and preferred features, as well as causal constraints which CFWoT guarantees will be respected. We demonstrate the performance of CFWoT against four baselines on several datasets and find that, despite not having access to a training dataset, CFWoT finds CFEs that make significantly fewer and significantly smaller changes to the input time-series. These properties make CFEs more actionable, as the magnitude of change required to alter an outcome is vastly reduced.
Rearrangement of three types of fused polycyclic hydrocarbons with carbon atoms 12–14 was conducted using chloroaluminate ionic liquid (IL) as acid catalyst. The hydrocarbons undergo quick configurational isomerization and then skeletal rearrangement toward alkyl-adamantanes. Although having different molecular structure, these hydrocarbons all lead to adamantanes substituted with 2–3 methyl and/or ethyl groups. The alkyl-adamantanes show high density, low freezing point and viscosity, and are very attractive as high-density fuels. Computation using density functional theory confirms that alkyl-adamantanes are thermodynamically favored and explains the experimental product distribution very well. On this basis, the reaction pathway of each hydrocarbon was illustrated. The reaction conditions including acidity (AlCl3 fraction), temperature, and IL dosage show significant effect on the rearrangement. Moreover, the distribution of alkyl-adamantanes can be tuned by adjusting the reaction conditions, allowing the fuels' properties to be fine-tuned. This work provides a versatile route to synthesizing alkyl-diamondoid fuels using simple chemical feedstocks in an effective way.
Cholangiocarcinoma (CCA) is a malignancy arising from multiple locations along the biliary tree, which is still lacking effective diagnostic biomarkers. The present study aimed to provide a comprehensive differential gene expression profile for the disease. The differentially expressed genes (DEGs) for CCA were explored in‑depth using a Gene Expression Omnibus (GEO) dataset, an internal cohort of clinical participants as well as <em>in vitro</em> experiments with the HUCCT1 cell line. Based on the GEO dataset, potential biomarker genes were proposed and the enriched biological processes as well as signaling pathways were further investigated. A protein‑protein interaction network of target genes was established. In the clinical specimens, the functions of the primary candidate, FBJ murine osteosarcoma viral oncogene homolog B (FOSB), were evaluated by reverse transcription‑quantitative (RT‑q)PCR and western blot analysis. A Cell Counting Kit‑8 (CCK‑8) assay was used for a functional study on FOSB. The results indicated that, compared with non‑tumor bile duct tissues, primary CCA samples had 676 differentially expressed genes, including 277 downregulated and 399 upregulated ones. Among these, HBD, FOSB, HBB, ITIH2, FCGBP, MT1JP, PIJR, SLC38A1, COL10A1 and MMP19 were determined to be the most significant DEGs. At the same time, upregulated genes were enriched in 'vasculature development' and 'IL‑17 signaling pathways'. Downregulated genes were enriched in 'extracellular matrix progress' and 'glucuronate signaling pathway'. The patients with CCA displayed decreased levels of hemoglobin. Compared with paracancerous tissues, CCA cancerous tissues exhibited increased RNA and protein expression levels of FOSB according to RT‑qPCR and western blot analysis, respectively. Furthermore, FOSB expression influenced the proliferation/viability of the CCA cell line HUCCT1. In conclusion, the present study suggested that the FOSB gene may serve as a primary biomarker candidate for CCA, providing a valuable reference for its clinical implementation.
