Large language models (LLMs) are often described as multilingual because they can understand and respond in many languages. However, speaking a language is not the same as reasoning within a culture. This distinction motivates a critical question: do LLMs truly conduct culture-aware reasoning? This paper presents a preliminary computational audit of cultural inclusivity in a creative writing task. We empirically examine whether LLMs act as culturally diverse creative partners or merely as cultural translators that leverage a dominant conceptual framework with localized expressions. Using a metaphor generation task spanning five cultural settings and several abstract concepts as a case study, we find that the model exhibits stereotyped metaphor usage for certain settings, as well as Western defaultism. These findings suggest that merely prompting an LLM with a cultural identity does not guarantee culturally grounded reasoning.
Thanos Polychronou, Lukáš Adam, Viktor Penchev
et al.
Wildlife re-identification aims to recognise individual animals by matching query images to a database of previously identified individuals, based on their fine-scale unique morphological characteristics. Current state-of-the-art models for multispecies re- identification are based on deep metric learning representing individual identities by fea- ture vectors in an embedding space, the similarity of which forms the basis for a fast automated identity retrieval. Yet very often, the discriminative information of individual wild animals gets significantly reduced due to the presence of several degradation factors in images, leading to reduced retrieval performance and limiting the downstream eco- logical studies. Here, starting by showing that the extent of this performance reduction greatly varies depending on the animal species (18 wild animal datasets), we introduce an augmented training framework for deep feature extractors, where we apply artificial but diverse degradations in images in the training set. We show that applying this augmented training only to a subset of individuals, leads to an overall increased re-identification performance, under the same type of degradations, even for individuals not seen during training. The introduction of diverse degradations during training leads to a gain of up to 8.5% Rank-1 accuracy to a dataset of real-world degraded animal images, selected using human re-ID expert annotations provided here for the first time. Our work is the first to systematically study image degradation in wildlife re-identification, while introducing all the necessary benchmarks, publicly available code and data, enabling further research on this topic.
Viacheslav Vasilev, Vladimir Arkhipkin, Julia Agafonova
et al.
Despite the fact that popular text-to-image generation models cope well with international and general cultural queries, they have a significant knowledge gap regarding individual cultures. This is due to the content of existing large training datasets collected on the Internet, which are predominantly based on Western European or American popular culture. Meanwhile, the lack of cultural adaptation of the model can lead to incorrect results, a decrease in the generation quality, and the spread of stereotypes and offensive content. In an effort to address this issue, we examine the concept of cultural code and recognize the critical importance of its understanding by modern image generation models, an issue that has not been sufficiently addressed in the research community to date. We propose the methodology for collecting and processing the data necessary to form a dataset based on the cultural code, in particular the Russian one. We explore how the collected data affects the quality of generations in the national domain and analyze the effectiveness of our approach using the Kandinsky 3.1 text-to-image model. Human evaluation results demonstrate an increase in the level of awareness of Russian culture in the model.
The human-centered word association test (WAT) serves as a cognitive proxy, revealing sociocultural variations through culturally shared semantic expectations and implicit linguistic patterns shaped by lived experiences. We extend this test into an LLM-adaptive, free-relation task to assess the alignment of large language models (LLMs) with cross-cultural cognition. To address culture preference, we propose CultureSteer, an innovative approach that moves beyond superficial cultural prompting by embedding cultural-specific semantic associations directly within the model's internal representation space. Experiments show that current LLMs exhibit significant bias toward Western (notably American) schemas at the word association level. In contrast, our model substantially improves cross-cultural alignment, capturing diverse semantic associations. Further validation on culture-sensitive downstream tasks confirms its efficacy in fostering cognitive alignment across cultures. This work contributes a novel methodological paradigm for enhancing cultural awareness in LLMs, advancing the development of more inclusive language technologies.
Multi-animal tracking is crucial for understanding animal ecology and behavior. However, it remains a challenging task due to variations in habitat, motion patterns, and species appearance. Traditional approaches typically require extensive model fine-tuning and heuristic design for each application scenario. In this work, we explore the potential of recent vision foundation models for zero-shot multi-animal tracking. By combining a Grounding Dino object detector with the Segment Anything Model 2 (SAM 2) tracker and carefully designed heuristics, we develop a tracking framework that can be applied to new datasets without any retraining or hyperparameter adaptation. Evaluations on ChimpAct, Bird Flock Tracking, AnimalTrack, and a subset of GMOT-40 demonstrate strong and consistent performance across diverse species and environments. The code is available at https://github.com/ecker-lab/SAM2-Animal-Tracking.
Emilie A Paterson, Carly I O’Malley, Dawn M Abney
et al.
Primates are important species for biomedical research and ensuring their good welfare is critical for research translatability and ethical responsibility. Systematic animal welfare assessments can support continuous programme improvements and build institutional awareness of areas requiring more attention. A multi-facility, collaborative project aimed to develop and implement a novel primate welfare assessment tool (PWAT) for use with research macaques. PWAT development involved: establishing an internal focus group of primate subject matter experts, identifying animal welfare categories and descriptors based on literature review, developing a preliminary tool, beta-testing the tool to ensure practicality and final consensus on descriptors, finalising the tool in a database with semi-automated data analysis, and delivering the tool to 13 sites across four countries. The tool uses input- and outcome-based measures from six categories: physical, behavioural, training, environmental, procedural, and culture of care. The final tool has 133 descriptors weighted based upon welfare impact, and is split into three forms for ease of use (room level, site level, and personnel interviews). The PWAT was trialled across facilities in March and September 2022 for benchmarking current macaque behavioural management programmes. The tool successfully distinguished strengths and challenges at the facility level and across sites. Following this benchmarking, the tool is being applied semi-annually to assess and monitor progress in behavioural management programmes. The development process of the PWAT demonstrates that evidence-based assessment tools can be developed through collaboration and consensus building, which are important for uptake and applicability, and ultimately for promoting global improvements in research macaque welfare.
Eric N. Ponnampalam, Eric N. Ponnampalam, Michelle Kearns
et al.
Optimising resource use efficiency in animal- agriculture-production systems is important for the economic, environmental, and social sustainability of food systems. Production of foods with increased health enhancing aspects can add value to the health and wellbeing of the population. However, enrichment of foods, especially meat with health enhancing fatty acids (HEFA) increases susceptibility to peroxidation, which adversely influences its shelf life, nutritional value and eating quality. The meat industry has been challenged to find sustainable strategies that enhance the fatty acid profile and antioxidant actions of meat while mitigating oxidative deterioration and spoilage. Currently, by-products or co-products from agricultural industries containing a balance of HEFA and antioxidant sources seem to be a sustainable strategy to overcome this challenge. However, HEFA and antioxidant enrichment processes are influenced by ruminal lipolysis and biohydrogenation, HEFA-antioxidant interactions in rumen ecosystems and muscle biofortification. A deep understanding of the performance of different agro-by-product-based HEFA and antioxidants and their application in current animal production systems is critical in developing HEFA-antioxidant co-supplementation strategies that would benefit modern consumers who desire nutritious, palatable, safe, healthy, affordable, and welfare friendly meat and processed meat products. The current review presents the latest developments regarding discovery and application of novel sources of health beneficial agro-by-product-based HEFA and antioxidants currently used in the production of HEFA-antioxidant enriched ruminant meats and highlights future research perspectives.
In recent years, there has been significant interest in creating 3D avatars and motions, driven by their diverse applications in areas like film-making, video games, AR/VR, and human-robot interaction. However, current efforts primarily concentrate on either generating the 3D avatar mesh alone or producing motion sequences, with integrating these two aspects proving to be a persistent challenge. Additionally, while avatar and motion generation predominantly target humans, extending these techniques to animals remains a significant challenge due to inadequate training data and methods. To bridge these gaps, our paper presents three key contributions. Firstly, we proposed a novel agent-based approach named Motion Avatar, which allows for the automatic generation of high-quality customizable human and animal avatars with motions through text queries. The method significantly advanced the progress in dynamic 3D character generation. Secondly, we introduced a LLM planner that coordinates both motion and avatar generation, which transforms a discriminative planning into a customizable Q&A fashion. Lastly, we presented an animal motion dataset named Zoo-300K, comprising approximately 300,000 text-motion pairs across 65 animal categories and its building pipeline ZooGen, which serves as a valuable resource for the community. See project website https://steve-zeyu-zhang.github.io/MotionAvatar/
Large language models (LLMs) are reported to be partial to certain cultures owing to the training data dominance from the English corpora. Since multilingual cultural data are often expensive to collect, existing efforts handle this by prompt engineering or culture-specific pre-training. However, they might overlook the knowledge deficiency of low-resource culture and require extensive computing resources. In this paper, we propose CultureLLM, a cost-effective solution to incorporate cultural differences into LLMs. CultureLLM adopts World Value Survey (WVS) as seed data and generates semantically equivalent training data via the proposed semantic data augmentation. Using only 50 seed samples from WVS with augmented data, we fine-tune culture-specific LLMs and one unified model (CultureLLM-One) for 9 cultures covering rich and low-resource languages. Extensive experiments on 60 culture-related datasets demonstrate that CultureLLM significantly outperforms various counterparts such as GPT-3.5 (by 8.1%) and Gemini Pro (by 9.5%) with comparable performance to GPT-4 or even better. Our human study shows that the generated samples are semantically equivalent to the original samples, providing an effective solution for LLMs augmentation. Code is released at https://github.com/Scarelette/CultureLLM.
This paper addresses the challenge of animal re-identification, an emerging field that shares similarities with person re-identification but presents unique complexities due to the diverse species, environments and poses. To facilitate research in this domain, we introduce OpenAnimals, a flexible and extensible codebase designed specifically for animal re-identification. We conduct a comprehensive study by revisiting several state-of-the-art person re-identification methods, including BoT, AGW, SBS, and MGN, and evaluate their effectiveness on animal re-identification benchmarks such as HyenaID, LeopardID, SeaTurtleID, and WhaleSharkID. Our findings reveal that while some techniques generalize well, many do not, underscoring the significant differences between the two tasks. To bridge this gap, we propose ARBase, a strong \textbf{Base} model tailored for \textbf{A}nimal \textbf{R}e-identification, which incorporates insights from extensive experiments and introduces simple yet effective animal-oriented designs. Experiments demonstrate that ARBase consistently outperforms existing baselines, achieving state-of-the-art performance across various benchmarks.
In livestock species, the monolayer of epithelial cells covering the digestive mucosa plays an essential role for nutrition and gut barrier function. However, research on farm animal intestinal epithelium has been hampered by the lack of appropriate in vitro models. Over the past decade, methods to culture livestock intestinal organoids have been developed in pig, bovine, rabbit, horse, sheep and chicken. Gut organoids from farm animals are obtained by seeding tissue-derived intestinal epithelial stem cells in a 3-dimensional culture environment reproducing in vitro the stem cell niche. These organoids can be generated rapidly within days and are formed by a monolayer of polarized epithelial cells containing the diverse differentiated epithelial progeny, recapitulating the original structure and function of the native epithelium. The phenotype of intestinal organoids is stable in long-term culture and reflects characteristics of the digestive segment of origin. Farm animal intestinal organoids can be amplified in vitro, cryopreserved and used for multiple experiments, allowing an efficient reduction of the use of live animals for experimentation. Most of the studies using livestock intestinal organoids were used to investigate host-microbe interactions at the epithelial surface, mainly focused on enteric infections with viruses, bacteria or parasites. Numerous other applications of farm animal intestinal organoids include studies on nutrient absorption, genome editing and bioactive compounds screening relevant for agricultural, veterinary and biomedical sciences. Further improvements of the methods used to culture intestinal organoids from farm animals are required to replicate more closely the intestinal tissue complexity, including the presence of non-epithelial cell types and of the gut microbiota. Harmonization of the methods used to culture livestock intestinal organoids will also be required to increase the reproducibility of the results obtained in these models. In this review, we summarize the methods used to generate and cryopreserve intestinal organoids in farm animals, present their phenotypes and discuss current and future applications of this innovative culture system of the digestive epithelium.
The present study's objective was to evaluate the inhibitory activity of the Probiotics Lactobacillus acidophilus and Bifidobacterium (obtained from the Agriculture Research Directorate, Ministry of Science and Technology, Iraq) and a suspension of a mixture between the two mentioned probiotics with two types of Gram-negative bacteria (Pseudomonas spp and Proteus spp) and one type of Gram-positive bacteria (Streptococcus spp) in vitro. The required tests were completed to verify the probiotics' purity, and the bacterial isolates used in the current investigation were assessed using biochemical assays and selected culture medium (culture and microscopic features). In addition, the inhibitory efficacy of the investigated Probiotics in different Gram positive and negative bacteria was evaluated by drug susceptibility testing (disc diffusion test as well as agar well diffusion test). Our data of the current study confirmed an excellent inhibitory activity of each Bifidobacterium (B) and the mixture of the two probiotics (MLB) via measuring the inhibition area, they had 25, 22mm, 28,-30 mm inhibition zone for Pseudomonas spp, 23, 25 mm, 26-27mm inhibition zone for Proteus species spp, and 22,20 mm, 33,29 mm inhibition quarter for Streptococcus species, by way of the usage of disc and agar well diffusion methods respectively. Where it was once weak inhibition activity of Lactobacillus acidophilus (L)on Pseudomonas spp, 0-3 mm and Streptococcus species 1-7 mm by the usage ofthe disc and agar well diffusion respectively. On the other, hand, Probiotic(Lactobacillus acidophilus) had available zone of inhibition on the Proteus sppbacteria, which were 24, 24 mm through the disc and agar well diffusion respectively. In conclusion: the Probiotics were found to have good and active inhibitory action on Gram-positive microorganism (Streptococcus) and gram-negative microorganism (Pseudomonas and Proteus) in vitro by way of using disc and agar well diffusion test, and the combination of the two probiotics MBL of present study, had more potent inhibitory action than each one of the studied separate probiotics.
Mostafa Abdollahi, Samad Lotfollahzadeh, Sara Shokrpoor
et al.
Abstract Six 100‐day‐old mixed‐breed lambs were examined in a farm with a semi‐intensive system due to neurologic signs. Cachexia, bilateral blindness, stupor, severe drowsiness and lethargy with left and right movements of the head and neck were recorded after awakening and stimulation. Lambs died 10 days after the onset of the clinical signs. The lambs were necropsied, and after routine parasitology, bacteriology and histopathology, the occurrence of acute coenurosis was confirmed due to finding multiple cystic structures in the brain tissue. All lambs of the herd were treated with albendazole (orally, 25 mg/kg, two doses with an interval of 14 days). All shepherd dogs were treated with popantel (orally, one tablet/10 kg, two doses with an interval of 14 days). The affected lambs died despite this treatment. No new case of the disease was observed after the initiation of control measures. The present study shows the importance of preventive measure against coenurosis in a semi‐intensive sheep farming system that includes implementing consistent parasite control programme in dogs being in contact with sheep.
ABSTRACT Brazilian chicken meat is exported to more than 150 countries and consumed by consumer markets that demand high quality and food safety, thus, requiring very strict control of pathogens present in food to guarantee these rigorous safety standards. This study evaluates the reports from the Salmonella spp. Control and Monitoring Program of the Brazilian Ministry of Agriculture, Livestock and Food Supply of seven slaughterhouses inspected by the Federal Inspection Service from the western region of Paraná state, Brazil, from March 2017 to February 2019. The broiler litter swab and carcass analyses revealed a Salmonella spp. positivity ratio of 5.9% (19/319) and 23.5% (75/319), respectively. The concomitant presence of Salmonella spp. in the broiler litter swab and chicken carcasses occurred in 58% of the positive samples. The most frequently isolated serovar in the carcasses was Salmonella Heidelberg (85.3%) followed by Salmonella spp. (10.6%). During slaughter, carcass positivity to Salmonella spp. was significantly different (p=0.047) between the first (19.6%) and the second (29.4%) shifts. The results alert for the possibility of carcass contamination during slaughtering and, therefore, more stringent hygiene measures between shifts must be implemented to mitigate carcass contamination.
Pixel-aligned Implicit Function (PIFu) effectively captures subtle variations in body shape within a low-dimensional space through extensive training with human 3D scans, its application to live animals presents formidable challenges due to the difficulty of obtaining animal cooperation for 3D scanning. To address this challenge, we propose the combination of two-stage supervised and self-supervised training to address the challenge of obtaining animal cooperation for 3D scanning. In the first stage, we leverage synthetic animal models for supervised learning. This allows the model to learn from a diverse set of virtual animal instances. In the second stage, we use 2D multi-view consistency as a self-supervised training method. This further enhances the model's ability to reconstruct accurate and realistic 3D shape and texture from largely available single-view images of real animals. The results of our study demonstrate that our approach outperforms state-of-the-art methods in both quantitative and qualitative aspects of bird 3D digitization. The source code is available at https://github.com/kuangzijian/drifu-for-animals.
Carbapenemase-producing Enterobacterales (CPE) are one of the most urgent threats to human healthcare globally. Our study aimed at optimizing the chromogenic agar method of screening for CPE with and without selective enrichment to isolate CPE from animal feces. The selective enrichment step did not result in any increased recovery of CPE from companion animals with NDM-5 E. coli, which suggests that enrichment broth may not be necessary for outbreak surveillance testing. Therefore, the selective enrichment step is optional, however, this method may not be generalizable to the detection of all types of CPE in fecal specimens from companion animals.
This study was performed according to FDA protocol to evaluate the developmental effects of carvedilol (P-glycoprotein inhibitor), methotrexate (P-glycoprotein substrate) and their combination at therapeutic doses on pregnant rats. Sixty Albino Wistar rats (40 female rats and 20 males) were allocated randomly into four groups orally administered 0.72 mg/kg carvedilol (Cv-treated group [TG]), 0.36 mg/kg methotrexate (MTX-TG), combined doses carvedilol+methotrexate (Cv+MTX-TG), and distilled water (control group) for 2 months in male and 2 weeks in female rats before mating and after copulation, then approval of pregnancy; dosing continued in female groups during pregnancy and lactation periods. Half of the animal groups were euthanized one day before parturition to study prenatal effects, while the other half left for parturition and lactation to study postnatal effect. The results of fertility index recorded in Cv-TG (71.43%), MTX-TG (42.46%) and Cv+MTX-TG (38.47%) was markedly lower than that in control (83.33%) group with lower gestation index was recorded in MTX-TG (80%) and Cv+MTX-TG (60%) than that in Cv-TG (100%) and the control group (100%). The result of resorbed and fetal death recorded a higher percent in Cv-MTX-TG in comparison with MTX-TG and Cv-TG; Cv-MTX-TG fetuses also recorded more anomalies, including hemorrhagic placenta, curved legs, and microcephaly during prenatal period. The postnatal effects showed that the Cv+MTX-TG group recorded a higher decrease in number of pups born, their weight, and increase in number of stillbirths in comparison with methotrexate followed by carvedilol groups in comparison with control group, while the result of viability index recorded (Cv-TG=98.15%, MTX-TG=93.93% and Cv+MTX-TG=76.19%) and lactation index (Cv-TG=77.36%, 83.87% and Cv+MTX-TG=75%). The postnatal anomalies were only recorded in Cv+MTX-TG included skull defect and ulceration, blindness, skin lesion, and alopecia in lactating pups. It is concluded that inhibition of P-gp by carvedilol might increase the placental passage and increase methotrexate concentration in fetal and pups’ tissue with consequence of increase toxic effect of methotrexate both in fetus and pups of Cv+MTX-TG group which might explain the present results of teratogenic study.
Abstract Arctigenin (ACT) is a novel anti‐inflammatory lignan extracted from Arctium lappa L, a herb commonly used in traditional Chinese herbal medicine. In this study, we investigated the molecular mechanism whereby ACT inhibits PCV2 infection‐induced proinflammatory cytokine production in vitro and in vivo. We observed that in PCV2 infection+ACT treated PK‐15 cells, proinflammatory cytokine production was significantly reduced, compared to the PCV2‐infected cells. The transfection and luciferase reporter assay confirmed that ACT suppressed NF‐κB signalling pathway activation following PCV2 infection in PK‐15 cells. Furthermore, western blotting demonstrated that ACT suppressed the NF‐κB signal pathway in PCV2 infection‐stimulated PK‐15 cells by inhibiting the translocation of p65 from the cytoplasm to the nucleus and IκBα phosphorylation. BALB/c mice were used as a model to evaluate the anti‐inflammatory effect of ACT in vivo. We found that the BALB/c mice inoculated with PCV2 infection + ACT treated showed a significant reduction of proinflammatory cytokine production in serum, lung and spleen tissue, compared to the PCV2‐infected mice. Western blotting confirmed that ACT suppressed the NF‐κB signal pathway in PCV2‐infected mice by inhibiting the translocation of p65 from the cytoplasm to the nucleus and IκBα phosphorylation in lung tissue. Our studies first demonstrate that ACT inhibits PCV2 infection‐induced proinflammatory cytokine production by suppressing the phosphorylation and nuclear translocation of NF‐κB in vitro and in vivo. These results will help further develop ACT as a Traditional Chinese herbal medicine remedy in the treatment of porcine circovirus‐associated diseases.
Gabrielle Cardoso Reis Fontan, Danielle Jordana Soares do Nascimento, Arthur Pompilio da Capela
et al.
Skincare is a growing concern of the population and the consumption of hydrolyzed collagen as a nutricosmetic is a trend. Its incorporation is low in dairy products but presents great potential due to the high consumption of this class of products. Thus, the objective of this work was to prepare a passion fruit-flavored Petit Suisse cheese with added collagen and evaluate the influence of information on the collagen benefits in the acceptance of the product. The product was prepared with 5% of hydrolyzed collagen and microbiological and physicochemical analyses were performed before the sensory evaluation. Women over 24 years old were invited to answer a questionnaire and participate in the acceptance test. It was observed that women over 30 years are concerned with the health of the skin, consume some type of product for this purpose and accept to pay higher values for products with this purpose. The younger women (24 to 30 years old) showed low concern and less interest in products with added collagen. The sensory evaluation indicated that the average score in the 2nd session (7.72 - with information), was significantly higher when compared to the 1st session, (7.24 - without information), indicating that there was an influence of the information in the product acceptance. It was concluded that product acceptance is positively affected by the information and has great potential for the market.
BackgroundA new, portable bedside coagulation monitor (VCM Vet) has provided a user-friendly, cartridge-based method to perform viscoelastic testing. However, the use of native whole blood limits the time to analyze the sample to minutes. The objective of this study is to assess whether citrated whole blood can be utilized with the cartridge-based system and whether the results are comparable to those of native whole blood. A secondary objective is to assess the viability of citrated whole blood results after up to 4 hours of resting.MethodsThe study population consisted of 10 healthy mixed breed dogs. Whole blood samples were collected via jugular venipuncture. Blood was immediately transferred to the VCM test cartridge for native whole blood control group analysis per the manufacturer's instructions, and the remainder was used to fill two 3.2% sodium citrate vacutainer tubes. Test group analysis was performed on samples from each tube concurrently after a rest period of 30 min (baseline), 2 h, and 4 h. Citrated whole blood samples were recalcified for analysis immediately prior to introduction into the test cartridge. Data was recorded for all reported parameters. Results from the citrate groups were compared to the control group and to the citrated baseline to assess for differences. Overall results were compared using mixed ANOVA models. Where found, specific differences were evaluated using Tukey's test. Within-sample variation was investigated and reported as median (range). A p < 0.05 was considered significant.ResultsSamples were obtained for a total of 10 control runs and 20 citrated whole blood runs. Comparison of controls to the citrated test groups revealed significant differences in CT (p < 0.001) and MCF (p < 0.002). There were no significant differences between test groups compared to citrated baselines for any parameter. Selected median coefficients of variation were 6.8% (0–68.8%) for CT, 2.4% (0–19.46%) for alpha angle, 3.2% (0–27.4%) for MCF, and 0% (0–16.3%) for 45-min LY45.ConclusionCitrated whole blood samples can be used with the VCM Vet device; however, new reference intervals for use with citrated whole blood will be required. Results using citrated whole blood samples are not significantly different from baseline after up to 4 h of resting.