Informa’s Vishal Nigam Vishal Nigam on Large Multimodal Models

Vishal Nigam, global data and science manager at Informa, discusses the evolution from large language models (LLMs) to large multimodal models (LMMs), which can understand images and, in future, videos. These could even empower robots to learn using on-board vision systems.

He says the challenge of hallucinations still exists and has the potential of worse outcomes when associated with hardware, but overall the levels of hallucinations is reducing over time.

Nigam adds that LMMs have the potential to transform the way business is performed, given the example of the finding that 52% of U.S. social media users already follow a virtual influencer. Similar technology will give rise to hyper-personalization of marketing.