Mental workload contributes significantly to human performance. Considerable research exists on workload assessment using methods like subjective measurement and performance measurement. Physiological parameters like Electroencephalography (EEG) and Electrocardiography (ECG) have been recently used for automatic objective workload assessment. A few challenges remain including high performance assessment and workload assessment in multiple dimensions like visual, cognitive, and fine motor. To address this, IAI and its collaborators, Old Dominion University and the University of Iowa, have been awarded a contract entitled, “SMOLT: Sensitive Mental Workload Assessment Enhanced with Multi-Task Learning” A key innovation of the SMOLT software tool for multi-dimensional workload assessment model is incorporating advanced Multi-Task Learning (MTL) theory and multimodal deep learning, which models the relatedness among the output tasks or workload in different dimensions, and among input signals or multimodal deep learning for better feature representations. Physiological signals, like EEG and ECG are preprocessed, and a library of features extracted and the most prominent selected. The features are classified with a basic deep learning model based on unlabeled data and limited labeled data. Multimodal deep learning helps learn shared feature representation among different sensing modalities. SMOLT uses prior research by IAI and its collaborators in cognitive state assessment using advanced learning techniques in multi-dimensions. A series of signal processing techniques, including pre-processing, wavelet neural networks-based EEG artifact removal, a feature library for EEG/ECG signals, advanced feature selection algorithms and machine learning algorithms for cognitive state classifications have been developed and implemented. These existing algorithms will be incorporated into SMOLT software for more accurate workload assessment.