Search for: All records

The agricultural sector, particularly animal production, faces numerous unprecedented challenges driven by climate change, resource depletion, and an ever-growing global demand for quality food. These challenges are further compounded by the increasing environmental impact of livestock farming, including greenhouse gas emissions, overuse of water and land resources, and the destruction of vital ecosystems. Ensuring the sustainability of animal production systems while mitigating the negative environmental impacts of these factors is essential for future global food security. As the demand for animal-derived products continues to rise, there is a pressing need for innovations that can enhance productivity without compromising environmental integrity or animal welfare. Artificial intelligence (AI) has emerged as a transformative technology with the potential to revolutionize the animal production industry. AI-driven solutions offer promising avenues for optimizing production efficiency, enhancing animal health and welfare, and reducing the environmental footprint of livestock farming. Machine learning, sensor technologies, and advanced data analytics are being increasingly utilized to monitor and predict various aspects of animal farming, such as feed efficiency, disease prevention, and climate resilience. These technologies enable farmers to make data-driven decisions, fostering more sustainable and environmentally responsible practices. This review examines the integration of AI into animal production systems, emphasizing its applications in climate change mitigation, resource management, and advancing sustainability. The discussion addresses how AI technologies can be utilized to improve productivity while minimizing environmental impact and enhancing animal welfare. Additionally, the paper outlines future opportunities, challenges, and potential barriers to integrating AI technologies into livestock farming, thereby ensuring long-term sustainability amid global challenges. 

This work has two contributions. First, we introduce novel provably Byzantine-resilient sample- and communication-efficient alternating gradient descent (GD) and minimization based algorithms for solving the federated low rank matrix completion (LRMC) problem. This involves learning a low rank (LR) matrix from a small subset of its entries. Second, we extend our ideas to show how a simple modification of our algorithms also provably solves two other partly-decoupled vertically federated LR matrix learn- ing problem, – LR column-wise sensing (LRCS), also referred to as multi-task linear representation learning, and its phaseless generalization, LR phase retrieval (LRPR). In all problems, we consider column-wise or vertical federation, i.e. each node observes a small subset of entries of a disjoint column sub-matrix of the entire LR matrix. 

The problem of hovering in flapping insects/hummingbirds, and potential bio-mimicry by micro aerial vehicles (MAVs), have been studied for decades by scientists and engineers. Said communities often study hovering in flapping systems as either an open-loop or closed-loop system to analyze stability and/or propose control designs. Recently, a fundamentally novel result has been achieved in the literature of the hovering problem. That is, hovering in flapping insects/hummingbirds can be characterized/mimicked as a stable, model-free, real-time extremum seeking control (ESC) feedback system. In this paper we aim at two contributions: (i) provide a novel open-loop, optimal control characterization of hovering; and (ii) compare the performance of closed-loop, real-time ESC in hovering vs. the provided open-loop, non-real-time optimal control in hovering. 

NLP research has increasingly focused on subjective tasks such as emotion analysis. However, existing emotion benchmarks suffer from two major shortcomings: (1) they largely rely on keyword-based emotion recognition, overlooking crucial cultural dimensions required for deeper emotion understanding, and (2) many are created by translating English-annotated data into other languages, leading to potentially unreliable evaluation. To address these issues, we introduce Cultural Lenses on Emotion (CuLEmo), the first benchmark designed to evaluate culture-aware emotion prediction across six languages: Amharic, Arabic, English, German, Hindi, and Spanish. CuLEmo comprises 400 crafted questions per language, each requiring nuanced cultural reasoning and understanding. We use this benchmark to evaluate several state-of-the-art LLMs on culture-aware emotion prediction and sentiment analysis tasks. Our findings reveal that (1) emotion conceptualizations vary significantly across languages and cultures, (2) LLMs performance likewise varies by language and cultural context, and (3) prompting in English with explicit country context often outperforms in-language prompts for culture-aware emotion and sentiment understanding. The dataset and evaluation code are publicly available.