A research scientist position is available immediately at the Institute for Infocomm Research (I2R), A*STAR, Singapore. We are looking for strong candidates in the field of Artificial intelligence and deep learning, with an interest in working on both fundamental algorithms in large-scale machine learning or deep learning, and also applying it to practical problems in vertical domains.
The position, based within the Program in Artificial Intelligence at I2R, entails working closely with a highly interdisciplinary team of researchers with training from leading universities like MIT, Stanford and CMU. The appointment will be based in Singapore for 2 years duration. Listed below are problems of interest:
Deep learning application in AME (Semiconductor domain): Explore state-of-the-art deep learning technologies in Advance Manufacturing Engineering, especially in semiconductor industries. Including, but not limited to, yield analysis and prediction, predictive maintenance (PdM), advanced DoE analysis and prediction, etc.
Multimodal Deep Learning: Towards effective learning from heterogeneous data sources
Currently, deep learning is applied in various areas such as visual computing, speech recognition, text mining, etc. In each area, the learning model is trained with large amounts of data in isolation. A multimodal learning approach from heterogeneous data sources, including video, speech, text, etc. is necessary to achieve breakthrough in performance.
Incorporating Knowledge Graphs into Deep Learning: Towards better understanding of images
Current deep learning systems today have no real understanding of images, with output typically a bunch of bounding boxes, learnt using a large number of label or sentence-level associations. Knowledge graphs like the MIT ConceptNet will be used to understand higher-level semantics and relationships between entities, to improve classification accuracy.