Join our virtual meetup to hear talks from experts on cutting-edge topics across AI, ML, and computer vision. View more CV events here.

Rumen acidosis silently costs the global dairy industry billions annually and compromises animal welfare, yet current detection methods remain invasive, delayed, and impractical at scale. Our lab has pioneered a fundamentally new approach: capturing and analyzing exhaled CO₂ and CH₄ gas emission patterns through synchronized RGB-thermal imaging, turning every breath into a diagnostic signal. We developed DualGasNet, a dual-stream deep learning architecture with cross-attention fusion that detects acidosis non-invasively and in real time, achieving state-of-the-art accuracy on a first-of-its-kind livestock gas emission dataset we constructed from scratch. To push toward explainable, farm-ready AI, we integrate vision-language models — CLIP and LLaVA-1.5 — enabling zero-shot diagnostic reasoning that bridges the gap between deep learning predictions and actionable veterinary insight. This talk will walk through the full pipeline from custom dataset creation to multimodal fusion to VLM-powered interpretation, offering the audience a compelling case study in how computer vision can solve high-impact, real-world problems outside traditional benchmarks.

Vision–language models (VLMs) have achieved remarkable performance by aligning images and text in a shared Euclidean space. However, Euclidean embeddings struggle to capture hierarchical structures inherent in multimodal data, such as conceptual taxonomies or fine-grained categories. We propose HVL (Hyperbolic Vision–Language), a geometry-grounded framework that maps images and text into a Poincaré manifold to induce hierarchy-aware representations. HVL leverages the exponential capacity of hyperbolic space to preserve semantic distances across multiple levels of abstraction, while an adaptive, entropy-driven entailment loss enforces hierarchical ordering between modalities. Extensive experiments on zero-shot classification and image–text retrieval benchmarks demonstrate that HVL consistently outperforms Euclidean baselines (e.g., CLIP) and Lorentz-based hyperbolic models (e.g., MERU), particularly in scenarios requiring fine-grained hierarchical understanding. These results highlight the importance of respecting geometric structure and establish hyperbolic embeddings as a principled foundation for hierarchical multimodal representation learning.

This talk will explore practical workflows for building, evaluating, and improving modern computer vision systems. We’ll dive into real-world approaches to dataset curation, model analysis, multimodal AI workflows, and production-ready vision pipelines using open-source technologies.

The session is designed for engineers, researchers, and AI practitioners looking to better understand how teams are developing and scaling computer vision applications today. Expect practical demos, technical insights, and discussions around the evolving AI tooling ecosystem.