Artificial intelligence is now entering our everyday lives, helping housewives with their chores with smart home technology, office workers with their workload with smart office technology, and healthcare workers with machines to help them identify health issues at record speeds.
With a strong focus on related algorithms, Pensees is currently at the forefront of international research in this field, and its algorithms have been crucial to industrial applications, generating maximum productivity with minimum effort.
Here at Pensees, we use transfer learning to migrate the knowledge learnt from one domain to another, to identify and solve similar problems, and improve the ubiquity of AI, producing stellar results in our face recognition and re-identificaiton technologies.
As data collection becomes more important than ever, costs to collect and analyse this data are also rising. Pensees uses unsupervised learning to reduce the cost of data collection in abnormal behaviour recognition. We also employ few-shot learning, to reduce excessive dependence on big data, opting instead to get quality learning effects in small sample trainings.
To make our systems accessible to all, we have also delved into multimodal deep learning, where we integrate voice, vision, text, sensor and other multi-modal sensing signals to accommodate all human responses, and fulfil our mission by providing AI as a service to all.