Similar Attributes Research Articles

TextRefiner: Internal Visual Feature as Efficient Refiner for Vision-Language Models Prompt Tuning

Despite the efficiency of prompt learning in transferring vision-language models (VLMs) to downstream tasks, existing methods mainly learn the prompts in a coarse-grained manner where the learned prompt vectors are shared across all categories. Consequently, the tailored prompts often fail to discern class-specific visual concepts, thereby hindering the transferred performance for classes that share similar or complex visual attributes. Recent advances mitigate this challenge by leveraging external knowledge from Large Language Models (LLMs) to furnish class descriptions, yet incurring notable inference costs. In this paper, we introduce TextRefiner, a plug-and-play method to refine the text prompts of existing methods by leveraging the internal knowledge of VLMs. Particularly, TextRefiner builds a novel local cache module to encapsulate fine-grained visual concepts derived from local tokens within the image branch. By aggregating and aligning the cached visual descriptions with the original output of the text branch, TextRefiner can efficiently refine and enrich the learned prompts from existing methods without relying on any external expertise. For example, it improves the performance of CoOp from 71.66% to 76.96% on 11 benchmarks, surpassing CoCoOp which introduced instance-wise feature for text prompts. Equipped with TextRefiner, PromptKD achieves state-of-the-art performance while keep inference efficient.

Physics-Guided Fair Graph Sampling for Water Temperature Prediction in River Networks

This work introduces a novel graph neural networks (GNNs)-based method to predict stream water temperature and reduce model bias across locations of different income and education levels. Traditional physics-based models often have limited accuracy because they are necessarily approximations of reality. Recently, there has been an increasing interest of using GNNs in modeling complex water dynamics in stream networks. Despite their promise in improving the accuracy, GNNs can bring additional model bias through the aggregation process, where node features are updated by aggregating neighboring nodes. The bias can be especially pronounced when nodes with similar sensitive attributes are frequently connected. We introduce a new method that leverages physical knowledge to represent the node influence in GNNs, and then utilizes physics-based influence to refine the selection and weights over the neighbors. The objective is to facilitate equitable treatment over different sensitive groups in the graph aggregation, which helps reduce spatial bias over locations, especially for those in underprivileged groups. The results on the Delaware River Basin demonstrate the effectiveness of the proposed method in preserving equitable performance across locations in different sensitive groups.

Finite Dose In Vitro Permeation Testing: Significance of Occluding the Donor compartment, a Case study.

The evaporation of the solvent induces significant changes in formulation, directly impacting its performance. The performance of topical products is determined by the interplay between their inherent quality attributes and the transformations that occur due to solvent evaporation when applied to the skin in clinically relevant doses. To accurately assess, in vivo performance, it is advisable to apply smaller doses to the skin and keep the donor compartment open to enable evaporation of solvents while carrying out in vitro permeation tests. This manuscript highlights the critical role of solvent evaporation in differentiating the performance of two compositionally distinct products. One gel formulation contained alcohol, while the other did not. Although both exhibited similar quality attributes, their drying profiles varied significantly. Permeation studies conducted with closed donor compartments (Evaporation-disabled (ED)) failed to reveal these differences. However, when the donor compartments were exposed to the atmosphere to allow evaporation (Evaporation-enabled (EN)), the performance differences between the two products became evident.

Paleohistological Analysis of “Terror Birds” (Phorusrhacidae, Brontornithidae): Paleobiological Inferences

Extinct flightless birds of the clades Phorusrhacidae and Brontornithidae are among the most characteristic elements of the South American Cenozoic faunas. Although the anatomy, systematics, paleoecology, and distribution of these birds have been analyzed in some detail, there is not a single work dealing with the osteohistology of these taxa. Herein, we present for the first time paleohistological data for phorusrhacids (Patagornis marshi) and brontornithids (Brontornis burmeisteri). A dense vascularized fibrolamellar matrix and an uninterrupted cortex lacking growth lines indicate similar metabolism and physiological attributes occurring in most extant birds, and they differ from the condition of insular taxa or species living in very stable habitats. This is congruent with recent analyses indicating that phorusrhacids were very active birds lacking insularity-related attributes. It also supports recent growing evidence indicating that these birds’ extinction may be related to climatic and environmental changes rather than being the result of competitive displacement by immigrant placental mammals.

Efficient Hamiltonian encoding algorithms for extracting quantum control mechanism as interfering pathway amplitudes in the Dyson series

Hamiltonian encoding is a methodology for revealing the mechanism behind the dynamics governing controlled quantum systems. In this paper, following Mitra and Rabitz \cite{abhra_1}, we define mechanism via pathways of eigenstates that describe the evolution of the system, where each pathway is associated with a complex-valued amplitude corresponding to a term in the Dyson series. The evolution of the system is determined by the constructive and destructive interference of these pathway amplitudes. Pathways with similar attributes can be grouped together into pathway classes. The amplitudes of pathway classes are computed by modulating the Hamiltonian matrix elements and decoding the subsequent evolution of the system rather than by direct computation of the individual terms in the Dyson series. The original implementation of Hamiltonian encoding was computationally intensive and became prohibitively expensive in large quantum systems. This paper presents two new encoding algorithms that calculate the amplitudes of pathway classes by using techniques from graph theory and algebraic topology to exploit patterns in the set of allowed transitions, greatly reducing the number of matrix elements that need to be modulated. These new algorithms provide an exponential decrease in both computation time and memory utilization with respect to the Hilbert space dimension of the system. To demonstrate the use of these techniques, they are applied to two illustrative state-to-state transition problems.

Costly exploration produces stereotypes with dimensions of warmth and competence.

Traditional explanations for stereotypes assume that they result from deficits in humans (ingroup-favoring motives, cognitive biases) or their environments (majority advantages, real group differences). An alternative explanation recently proposed that stereotypes can emerge when exploration is costly. Even optimal decision makers in an ideal environment can inadvertently form incorrect impressions from arbitrary encounters. However, all these existing theories essentially describe shortcuts that fail to explain the multidimensionality of stereotypes. Stereotypes of social groups have a canonical multidimensional structure, organized along dimensions of warmth and competence. We show that these dimensions and the associated stereotypes can result from feature-based exploration: When individuals make self-interested decisions based on past experiences in an environment where exploring new options carries an implicit cost and when these options share similar attributes, they are more likely to separate groups along multiple dimensions. We formalize this theory via the contextual multiarmed bandit problem, use the resulting model to generate testable predictions, and evaluate those predictions against human behavior. We evaluate this process in incentivized decisions involving as many as 20 real jobs and successfully recover the classic dimensions of warmth and competence. Further experiments show that intervening on the cost of exploration effectively mitigates bias, further demonstrating that exploration cost per se is the operating variable. Future diversity interventions may consider how to reduce exploration cost, in ways that parallel our manipulations. (PsycInfo Database Record (c) 2025 APA, all rights reserved).

Visual target tracking based on fractional – order bidirectional hybrid attentional feature fusion

This paper mainly designs and implements target tracking based on fractional – order feature fusion to solve tracking drift and tracking box jumping in complex scenes. Firstly, the starting point and overall structure of the model design were introduced. Secondly, in response to the low information utilization of the feature extraction network in the target tracking framework at the local scale of the target, a fractional – order node function based attention CNN hybrid attention extraction module was proposed to improve the robustness of target tracking. Finally, overall performance was quantitatively and qualitatively evaluated on multiple evaluation datasets with various advanced trackers. The results showed that the algorithm proposed in this paper had a high tracking advantage under severe changes in scale morphology, dynamic blurring, and similar interference attributes.

Node Classification in Networks via Simplicial Interactions.

In the node classification task, it is natural to presume that densely connected nodes tend to exhibit similar attributes. Given this, it is crucial to first define what constitutes a dense connection and to develop a reliable mathematical tool for assessing node cohesiveness. In this article, we propose a probability-based objective function for semi-supervised node classification that takes advantage of higher order networks' capabilities. The proposed function reflects the philosophy aligned with the intuition behind classifying within higher order networks, as it is designed to reduce the likelihood of nodes interconnected through higher order networks bearing different labels. In addition, we propose the stochastic block tensor model (SBTM) as a graph generation model designed specifically to address a significant limitation of the traditional stochastic block model (SBM), which does not adequately represent the distribution of higher order structures in real networks. We evaluate the objective function using networks generated by the SBTM, which include both balanced and imbalanced scenarios. Furthermore, we present an approach that integrates the objective function with graph neural network (GNN)-based semi-supervised node classification methodologies, aiming for additional performance gains. Our results demonstrate that in challenging classification scenarios-characterized by a low probability of homo-connections, a high probability of hetero-connections, and limited prior node information-models based on the higher order network outperform pairwise interaction-based models. Furthermore, the experimental results suggest that integrating our proposed objective function with existing GNN-based node classification approaches enhances the classification performance by efficiently learning higher order structures distributed in the network.

Cluster analysis of social determinants of health and HIV/AIDS knowledge among Peruvian youths using Kohonen’s self-organized maps: a data-exploration study based on a Demographic and health survey

ABSTRACT Background Human immunodeficiency virus (HIV) and acquired immunodeficiency syndrome (AIDS) have evolved into a global development burden, with nearly 40 million infections and 25 million deaths. Compared to other age groups, youth have increased risks of contracting the disease due to social and health structural factors; thus, additional efforts are needed to effectively tackle the challenges associated with this age group. Epidemiological studies employing unsupervised learning techniques are essential for shaping public health policies. Objective This study aimed to describe the Peruvian youth population based on their sociodemographic, health, and economic characteristics using an unsupervised learning approach through the development of a neural network model based on Kohonen’s self-organizing maps (SOMs), allowing the identification of social profiles in the study population. Methods This quantitative study used data from the 2019 Peruvian Demographic and Family Health Survey. An SOM network model for clustering individuals with similar attributes and clustering prototype vectors based on the agglomerative hierarchical clustering (AHC) method and their visualization on an SOM was applied to the study sample. Results Clustering of prototype vectors yielded four clusters, each of which represented a profile of Peruvian youths based on their knowledge of HIV/AIDS and structural health determinants. Conclusions Kohonen’s neural networks allowed the identification of patterns and behaviors among youths in Peru, quantifying and characterizing the four social clusters regarding HIV/AIDS and their social determinants. Kohonen’s maps may benefit healthcare professionals and policymakers by offering a useful method for tailoring interventions and policies based on the detected profiles, thereby enhancing the visibility of these focal points at the national level.

Exploring the Common Ground: A Comparative Analysis of Zoroastrianism and Islam

Islam and Zoroastrianism share remarkable similarities in their theological and eschatological beliefs. They both believe in the oneness of God, with different names (Ahura Mazda for Zoroastrians and Allah for Muslims) but similar attributes of omnipotence, omniscience, and omnipresence. Both religions acknowledge archangels, with Zoroastrianism recognizing six and Islam recognizing four, along with other minor angels. Both religions consider humans as the pinnacle of creation and emphasize the establishment of justice and truth on earth. They share beliefs about the nature of the universe as a creation of God, and the role of prophets as messengers of God. Prayer is considered important in both religions, with similar practices of five daily prayers (Salat in Islam, Gah Worship in Zoroastrianism) and similar beliefs about the afterlife, including the immortality of the soul, resurrection of the dead, balance of judgment, heaven and hell, and the existence of a bridge to be crossed after death (Chinvat Bridge in Zoroastrianism, Al-Siraat in Islam). Ethical codes emphasizing purity of mind and body are also similar in both religions. Thus, this article argues, using intertextuality theory, that Islam is not a new beginning but rather a continuation of earlier revelations, such as Zoroastrianism. This perspective further suggests that Zoroastrianism can be categorized within the Abrahamic traditions. JASBH, Vol. 69(2), 2024, pp. 223-237

Fourier analysis of signal dependent noise images

An archetype signal dependent noise (SDN) model is a component used in analyzing images or signals acquired from different technologies. This model-component may share properties with stationary normal white noise (WN). Measurements from WN images were used as standards for making comparisons with SDN in both the image domain (ID) and Fourier domain (FD). The ID wavelet expansion was applied to WN images (n = 1000). Orthogonality conditions were used to parametrically model the variance decomposition, as described in both domains. FD components were investigated with probability density function modeling and summarized measures. SDN images were constructed by multiplying both simulated and clinical mammograms (both with n = 1000) by WN. The variance decomposition for both WN and SDN decreases exponentially as a parametric function of the ID expansion level; expansion image variances for both types of noise were captured similarly in the Fourier plane corresponding with the ID parametric model. The Fourier transform of WN has a uniform power spectrum distributed exponentially; SDN has similar attributes. Fourier inversion of the lag-autocorrelation performed in the FD produced a statistical estimation of the SDN’s image factor. These findings are counterintuitive as SDN can be nonstationary in the ID but have stationary attributes in the FD.

Detection of Human Immunodeficiency Virus in Saliva Using a Nickel-Based Electrocatalyst

Acquired immunodeficiency syndrome (AIDS) is an ongoing and chronic condition caused by the human immunodeficiency virus (HIV). Early detection is one way to minimize the spread of AIDS. Current methods cannot reliably detect HIV in saliva samples due to the low concentration of viral particles in oral fluid. A simple, label-free, rapid response and unprecedented nickel-based electrochemical biosensor was developed to detect HIV in saliva using a NiOOH electrocatalyst. The current response difference at an applied potential vs a reference electrode in the presence of HIV was the trait of this sensor. It is hypothesized that the HIV p24 protein blocks the active surface area of the sensor, causing the current response to drop, which increases as the concentration of HIV increases in saliva. Comparative analyses using proteins with varying chemical moieties demonstrated that the sensor’s response was triggered by the functional group’s presence in the HIV p24 protein. The sensor’s limit of detection in saliva is ∼48.40 × 10–5 TCID50 ml−1 HIV culture fluid. The sensor was utilized to successfully distinguish the current response of HIV in saliva from twelve different pathogens. Finally, the sensor replicated similar attributes with HIV-positive patient’s saliva during a clinical study, which makes it a state-of-the-art non-invasive rapid response HIV sensor.

EPCO-33. OMICON: A CLOUD PLATFORM FOR FAIR RESEARCH ON GENE COEXPRESSION NETWORKS IN NORMAL HUMAN BRAIN AND BRAIN TUMORS

Abstract Genome-wide coexpression analysis of bulk human brain samples is a powerful approach for identifying highly reproducible transcriptional signatures of cell types and cell states, since it can survey huge numbers of individuals, cells, and transcripts. However, it is difficult to optimize gene coexpression network construction, compare gene coexpression modules from independent datasets, or even locate gene expression datasets with similar attributes. To address these challenges, we have developed the OMICON platform (theomicon.ucsf.edu) to promote FAIR (Findable, Accessible, Interoperable, Reusable) research practices involving human brain gene coexpression networks. OMICON contains structured gene expression data for >17K bulk human brain samples (~10K normal and ~7K malignant glioma), which were collected from diverse public repositories and consortia (e.g., GEO, TCGA, CGGA, CPTAC, REMBRANDT, IvyGAP, GTEx, NABEC, the Allen Institute). To promote interoperability, we have standardized metadata for hundreds of dataset and sample attributes, including single-nucleotide and copy-number mutations. For each dataset, we have used the FindModules R function to construct dozens of gene coexpression networks by iterating over module detection parameters. These efforts have identified ~100K gene coexpression modules, which have been extensively characterized via enrichment analysis with thousands of curated gene sets. All datasets, metadata, gene coexpression networks, enrichment results, and analysis steps can be browsed with an interactive workflow visualization tool, which promotes accessibility, reusability, and reproducibility by maintaining complete data provenance with unique identifiers. To promote findability, we have built an advanced search engine to identify datasets, samples, modules, and more, by filtering standardized metadata (e.g., find gene coexpression modules in human glioblastoma datasets that are significantly enriched with markers of T-cells). A detailed description of OMICON’s functionality is described on our Help page: theomicon.ucsf.edu/home/help. Through this functionality, OMICON seeks to build community and focus therapeutic efforts around reproducible analyses of transcriptional variation in normal human brain and brain tumors.

Inertial active Ornstein–Uhlenbeck particle in a non-linear velocity dependent friction

Inertial active Ornstein–Uhlenbeck particle in a non-linear velocity dependent friction

Handwritten Collection of Irkutsk Archpriest Fortunat Petukhov: The Reconstruction of the Composition and Problems of Search

The purpose of the article is to present the handwritten collection of the Irkutsk archpriest Fortunat Ivanovich Petukhov (1789–1872), formed in the middle of the 19th century. The author has made an attempt 1) to reconstruct the composition of the manuscripts and documents of the collection, 2) to trace the ways of its formation, 3) to identify the features of the selection of manuscripts and to trace the degree of influence on the fate of the collection of the first presentation in the work by Moscow scholar N. P. Bocharov. Textual analysis of a number of works mentioned in the list has made it possible for other authors to attribute some of the texts presented by N. P. Bocharov as manuscripts of his collection, and not as the author’s works of Fr Fortunatus. We are talking about the well-known work of the Irkutsk local historian N. V. Semivsky, used for the autobiographical notes by F. I. Petukhov and as an integral part of other works by the archpriest, as well as about the manuscript of a much lesser-known work by another Irkutsk priest, Fr Nikifor Parnyakov. This gives reason to assume the possibility of a similar attribution of other works indicated by N. P. Bocharov with the name of F. I. Petukhov. Several manuscripts from the collection of F. I. Petukhov were published by the biographer as part of his book. The location of the collection of Archpriest Fortunatus Petukhov has not been found; the article suggests directions for searching for a possible repository – materials from the archival funds of Moscow monasteries, connected with the life and activities of people from Siberia – Zlatoust, Sretensky, etc., as well as searches for the personal archives of Fr Fortunatus’ descendants.

Distinct Ventral Hippocampus Network Properties inDissociated Cultures.

Extensive research has been focused in the past century on structural, physiological, and molecular attributes of the hippocampus. This interest was created by the unique involvement of the hippocampus in cognitive and affective functions of the brain. Functional analysis revealed that the hippocampus has divergent properties along its axial dimension to the extent that the dorsal sector (dorsal hippocampus, DH) has different connections with the rest of the brain than those of the ventral sector (VH). Still, longitudinal pathways connect the DH with the VH and dampen the functional differences between the sectors. To be able to identify the intrinsic functional difference between the DH and VH, we produced dissociated monolayer cultures from prenatal DH and VH and examined their properties at 10-20 days after plating by imaging the spontaneous activity of the network using Fluo-2 AM, a calcium indicator. Surprisingly, while DH and VH sectors produced dissociated cultures with similar morphological attributes, VH cultures were more active spontaneously than DH cultures. Furthermore, when stimulated to produce action potentials, VH neurons triggered network bursts in postsynaptic neurons more often than DH cultures. Finally, in both DH and VH cultures, electrical stimulation of single cells produced network bursts in response to a burst of action potentials rather than to single spikes. These experiments indicate that even in dissociated cultures, neurons of the VH are more excitable and sensitive to electrical stimulation than DH; hence, they are more likely to generate network bursts and epileptic seizures, as suggested for invivo brains.

An Analysis of the Effect of Cross-Enrollment between High School and University on the Labor Market Performances

This study analyses the effect of cross-enrollment between high school and college on the college GPA as well as labor market performances. It uses a three-year(2017-2019) survey of the Graduates Occupational Mobility Survey(GOMS) and sets up the comparison group with similar attributes of the treated group based on the propensity score matching method. The results show that liberal arts students at high school who enrolled in science majors acquired a lower GPA at college than the comparison groups, but they were advantageous in employment. Similar results were found among science students who enrolled in liberal arts majors, but the results were not statistically significant. In addition, no meaningful differences were found in the probability of getting permanent job employments nor in earning levels. The results imply that cross-enrollment decisions do not seem to be a mere strategic behavior when it comes to college entrance. Therefore, it should to be encouraged to exploit the advantages of integrated learning rather than to be discouraged.

Discrete Choice Experiment in Evaluating Preferences for Physical Activity: A Mixed-Methods Study of Attributes and Their Levels.

Discrete choice experiment (DCE) has become a leading method for evaluating health sector preferences. Despite its common use in evaluating physical activity (PA) preference, there is a lack of consensus on the optimal attribute and level, and this step often lacks rigorous documentation. This study aimed to identify suitable attributes and levels for a PA program for future application in DCE studies. This study used a mixed-method approach. Initially, a rapid review was performed using databases and search engines. The findings were then used for a 2-stage semi-structured interview for attribute addition and scoring. Finally, top-scoring attributes were selected, and a small expert panel determined their associated levels. In our rapid review, from 8144 titles according to search strategy, 10 were selected for data extraction. After content analysis and integration of similar attributes, 12 were identified and 4 more were added from interviews. Using the opinions of the expert panel, the top 5 attributes were selected and a total of 22 levels were determined for these 5 attributes. The selected attributes included monthly cost, companions, distance, PA time, and PA type, with total scores of 150, 149, 147, 144, and 123, respectively. Our study identified 5 attributes and 22 corresponding levels as effective tools for measuring PA preferences using a DCE approach. As this is a pioneering study, further comprehensive research is recommended for improved outcomes.

Democrats and Republicans choose solar panels in very similar ways.

Appealing to individuals' social identity is a powerful form of social influence, capable of changing the way people process information, the information they think about, and how they evaluate other individuals. The purpose of this study is to explore the idea that Democrat and Republican environmental norms may impact the attributes and strategies partisans use when choosing whether to have solar panels on a house. An online study with N = 363 participants was conducted to examine these possible effects through multi-attribute decision making, applying predefined decision process models to participant behavior to test which attribute-based models best describe participants' decision making. A choice task was combined with an experimental manipulation of political affiliation salience to examine whether the norms of political groups would have influence on decision behavior. Results of the study show remarkable similarities between political parties in their strategies for choosing solar panels. Members of both political parties appeared to use similar strategies and similar attributes for the formation of their decisions. Recommendations are made that science communicators and policy makers avoid polarizing language so as not to create unnecessary polarization where ideological gaps may not currently exist.

A Multivariate Fuzzy Weighted K-Modes Algorithm with Probabilistic Distance for Categorical Data

Data clustering is a data mining approach that assigns similar data to the same group. Traditionally, cluster similarity considers all attributes equally, but in real-world applications, some attributes may be more important than others. Therefore, this study proposes an algorithm that utilizes multivariate fuzzy weighting to demonstrate the varying importance of each attribute, using a Gini impurity measure for weight assignment. Additionally, the proposed algorithm implements probabilistic distance to reduce sensitivity to noise. Probabilistic distance offers more detailed information and better interpretation than Hamming distance, which ignores attribute positions. Probabilistic distance utilizes information about the attribute’s position within and between clusters. This enhances clustering performance by creating clusters with more similar attributes. Therefore, the proposed Multivariate Fuzzy Weighted K-Modes with Probabilistic Distance for Categorical Data (MFWKM-PD) algorithm, based on the multivariate fuzzy K-modes algorithm, not only considers detailed membership calculations but also considers the varying contributions of attributes and their positions in distance calculation. This study evaluated the proposed MFWKM-PD using several benchmark datasets. The experiments validated that the proposed MFWKM-PD shows promising results compared to other algorithms in terms of accuracy, NMI, and ARI.