Skip to content
GCC AI Research

Search

Results for "MAviS"

Fine-grained species recognition with MAviS: a new dataset, benchmark, and model

MBZUAI ·

MBZUAI researchers have developed MAviS, a new multimodal dataset, benchmark, and chatbot for fine-grained bird species recognition. MAviS includes images, audio, and text to help models identify subtle differences between species, especially rare and regional varieties. The related study was presented at EMNLP 2025 and selected as a "Senior Area Chair Highlight". Why it matters: This work addresses a key limitation in AI's ability to support biodiversity conservation and ecological monitoring in the region and globally.

MBZUAI researchers earn high-profile honors at EMNLP

MBZUAI ·

MBZUAI researchers received high honors at EMNLP 2025 for two research papers, placing them in the top 2% of accepted work. One paper, MAviS, is a multimodal AI system that identifies bird species by combining images, sounds, and text. The other award-winning paper focuses on uncertainty in LLM-as-a-Judge. Why it matters: The recognition highlights MBZUAI's growing influence in NLP and multimodal AI research, particularly in domain-specific applications like biodiversity conservation.

MBZUAI hosts 3rd Mailis program to explore AI’s role in shaping future of work

MBZUAI ·

MBZUAI hosted the third edition of its Mailis program, focusing on AI's role in redefining the workplace and strengthening the UAE’s innovation ecosystem. The session brought together entrepreneurs, policymakers, and innovators from Abu Dhabi’s government, business, and startup sectors to discuss AI integration. Leaders from organizations like Khalifa Fund and Hub71 shared updates on AI initiatives, while MBZUAI showcased AI tools that can act as a co-founder. Why it matters: This event highlights the UAE's commitment to fostering AI adoption and collaboration across sectors, positioning the country as a leader in responsible AI innovation.

The ETH-MAV Team in the MBZ International Robotics Challenge

arXiv ·

The paper details the hardware and software systems of ETH Zurich's Micro Aerial Vehicles (MAVs) used in the 2017 Mohamed Bin Zayed International Robotics Challenge (MBZIRC). The team integrated computer vision, sensor fusion, and control to develop autonomous outdoor platforms. They achieved second place in Challenge 3 and the Grand Challenge, demonstrating autonomous landing in under a minute and a 90%+ visual servoing success rate for object pickups. Why it matters: The work highlights the advanced state of robotics research and development showcased at the MBZIRC, contributing to the growth of autonomous systems in the region.

MATRIX: Multimodal Agent Tuning for Robust Tool-Use Reasoning

arXiv ·

Researchers introduce MATRIX, a vision-centric agent tuning framework for robust tool-use reasoning in VLMs. The framework includes M-TRACE, a dataset of 28.5K multimodal tasks with 177K verified trajectories, and Pref-X, a set of 11K automatically generated preference pairs. Experiments show MATRIX consistently outperforms open- and closed-source VLMs across three benchmarks.

MBZUAI and Abu Dhabi Music & Arts Foundation explore AI’s Impact on the Arts at 4th Mailis

MBZUAI ·

MBZUAI and the Abu Dhabi Music and Arts Foundation (ADMAF) hosted the fourth edition of Mailis, focusing on AI's impact on the arts. The event, part of ADMAF’s Riwaq Al Fikr initiative, featured discussions on AI's role in music, visual art, and creative expression. A panel including MBZUAI's Gus Xia explored AI's creative potential in filmmaking and music. Why it matters: This collaboration highlights the growing interest in exploring the intersection of AI and culture in the UAE, potentially fostering new forms of artistic expression and cultural preservation.