Samsung R&D Institute Poland (SRPOL) was recognized as one of the leading teams at the Detection and Classification of Acoustic Scenes and Events (DCASE) 2022 challenge, held by the Institute of Electrical and Electronics Engineers (IEEE), which aims to use state-of-the-art artificial intelligence (AI) technology to understand and interpret audio signals.

SRPOL engineers with previous experience in the DCASE challenge (placing first in two tasks in 2019 and second in 2019 and 2020) set their focus on “Task 6B: Language-Based Audio Retrieval,” ranking third. This task seeks to evaluate methods where a retrieval system takes a free-form textual description as input and sorts audio signals in a fixed data set based on their match to the given description.

DCASE’s Language-Based Audio Retrieval fits flawlessly in the current strategy of SRPOL’s AI Team, which is renowned for projects and deployments in the Natural Language Processing (NLP) and Audio field. SRPOL utilizes its experiences and expertise daily to contribute to a better Bixby, provide better machine translations, and develop several other applications of NLP techniques (recent news on our last achievement in the NLP field).

Regarding Audio, SRPOL specializes in Sound Recognition, Sound Sources Separation, and Deep Signal Processing. Currently, SRPOL’s AI Team combines core technologies—NLP and Sound Recognition—and merges them into one product. From the user perspective, this opens new applications or extensions of existing ones.

Currently, users may browse video galleries based on fixed tags created through audio content analysis using Sound Recognition technology. However, merely browsing through fixed lists of tags is not as convincing. This is where NLP enters the scene and completes the picture. With combined Sound Recognition and NLP technologies, users could browse their content based on captions in natural language.