Exploring Analogical Reasoning through Iconicity in Sign Language and Gestures

The team at Pontificia Universidad Católica Del Peru led by Dr. Rodríguez Mondoñedo will review and explore reported iconicity criteria in Peruvian Sign Language (LSP) and American Sign Language (ASL) and works on other sign languages and non-verbal communication. First, they will define a list of words they will record from Peruvian signers and hearing people. Then, they will confirm the selection and iconicity level of the signs with Peruvian signer consultants. This list will be used in both the Peruvian and American teams to create a dataset of videos of a specific sign and variance aligned to images that correspond with the representation of that word. Then, they will report their findings on iconicity in two research papers and work with the CS Research Assistant to analyze their theories through computational methods such as skeleton-based datasets. Skeleton-based datasets are key landmarks identified in a person’s body, such as in the face, hands, fingers, and pose. Finally, they will also further analyze the results provided by the machine learning model that extends the sign language recognition representation to recognize other sign language and gestures performed by people who do not know sign language. The final results obtained with the machine learning models are expected to be analyzed by the Pontificia Universidad Católica Del Peru team and help interpret new insights between machine learning, analogical reasoning, sign languages, and non-verbal communication. These findings should be submitted to a venue or conference where this interdisciplinary work can be appreciated