Overview

In recent years, substantial advancements have been made across various domains of artificial intelligence, including vision and language models, audio processing, and robotics. However, these developments have largely occurred in isolation within their respective fields. While multimodal learning has garnered increasing attention, current approaches predominantly focus on a limited set of sensory modalities, namely vision, text, and occasionally audio. This restricted scope fails to encompass the full range of human sensory experience. A critical gap in contemporary AI research lies in the integration of diverse sensory modalities—extending beyond the conventional vision-language pairings—towards a more comprehensive, human-like understanding of the world. Human perception is inherently multisensory, seamlessly integrating inputs from touch, smell, taste, and interoception, alongside external senses such as sight and hearing. These additional modalities are not supplementary; rather, they are integral to how we interpret and engage with the world. Furthermore, the neuroscientific foundations of perception, as revealed through neural recordings like EEG, fMRI, and MEG, offer a valuable, yet underexplored, perspective for grounding AI systems in biologically-inspired mechanisms.

The goal of this workshop is to catalyze a paradigm shift in the AI community’s approach to perception: from a narrowly multimodal perspective to a truly multisensory one. We seek to highlight recent advancements in specialized domains, including computational olfaction, haptic learning, gustatory modeling, and neuro-symbolic integration, while bringing together researchers from diverse fields who may otherwise operate in isolation. In doing so, we aim to facilitate the cross-pollination of ideas and foster the development of AI systems that more closely mirror the richness and complexity of human perception.

Important Dates

Submission deadline: 18 August, 2025
Notification of acceptance: 15 September, 2025
Camera-ready submission: 22 September, 2025
Author registration: Coming soon through BMVC registration
Workshop date: 27 November 2025, Sheffield, UK

Keynote Speakers

Coming soon

Call for Papers

We welcome research contributions on Multisensory Intelligence and Multimodal Learning, including (but not limited to) the following topics:

• Multisensory data collection and benchmarking: touch (tactile), smell (olfactory), and taste (gustatory), paired with visual and language associations;
• Computational modeling of underexplored senses: representation learning, multimodal large language models (MLLMs);
• Alignment between sensory data and human perception: visual-tactile learning, multimodal brain encoding and decoding, linguistic-olfactory/gustatory learning;
• Sensory fusion and integration: handling noisy and heterogeneous data, and integrating with artificial agents;
• Multisensory applications: robotics, assistive technology, healthcare, and AR/VR;
• Ethical and design considerations for multisensory intelligent systems with human-like perception.

Format: Submission papers (.pdf format) must use the BMVC 2025 Author Kit for LaTeX/Word Zip file and be anonymized and follow BMVC 2025 author instructions. The review process will be double-blind.

Submission Site: Coming soon
Submission Due (both Paper and Supplementary Material): 18 August, 2025