Efficient Transformations in Deep Learning Convolutional Neural Networks

This study investigates the integration of signal processing transformations—Fast Fourier Transform (FFT), Walsh-Hadamard Transform (WHT), and Discrete Cosine Transform (DCT) within the ResNet50 convolutional neural network (CNN) model for image classification. The primary objective is to assess the trade-offs between computational efficiency, energy consumption, and classification accuracy during training and inference. Using the CIFAR-100 dataset (100 classes, 60,000 images), experiments demonstrated that incorporating WHT significantly reduced energy consumption while improving accuracy. Specifically, a baseline ResNet50 model achieved a testing accuracy of 66%, consuming an average of 25,606 kJ per model. In contrast, a modified ResNet50 incorporating WHT in the early convolutional layers achieved 74% accuracy, and an enhanced version with WHT applied to both early and late layers achieved 79% accuracy, with an average energy consumption of only 39 kJ per model. These results demonstrate the potential of WHT as a highly efficient and effective approach for energy-constrained CNN applications.

PharMe: A Pharmaceutical Informed LLM

This study explored the application of Medical LLaMA-3-8B, a large language model pre-trained on the MIMIC-III dataset, as a resource tool for assisting healthcare providers in drug and treatment selection based on patient diagnoses. The model was fine-tuned on data sourced (324 Updates for 2024) from Drugs@FDA, specifically the FDA’s Purple Book database, which provides comprehensive information on approved drugs, novel treatments, and biosimilars.

To ensure the model remained current with the latest updates, an automated workflow was developed using Apache Airflow. This workflow facilitated periodic data pulls from the FDA database, processed and formatted the retrieved information, and incorporated it into the fine-tuning loop.

The resulting fine-tuned model, termed PharMe, demonstrated the ability to provide valuable recommendations when prompted with medical conditions, achieving a perplexity score on average of 17.4, compared to GPT-Neo’s score of 21.7. PharMe not only suggested commonly prescribed treatments but also identified the latest advancements, including novel therapies, and biosimilars.

Hierarchical Voting-Based Feature Selection and Ensemble Learning Model Scheme for Glioma Grading with Clinical and Molecular Characteristics and Experiments with PCA and SMOTE Techniques

NJIT Senior Design Showcase Runner-up

Designing ultra-wide band (UWB) radar and, integration tool incorporates a range of components, including a transmitter, receiver, and an UWB pulse generator. It uses millimeter wave-length (mmWave) radiation, part of the electromagnetic spectrum with wavelengths typically in the range of 8 GHz to 20GHz, has specific properties when it comes to penetration through walls and other solid objects. With utilizing the Doppler effect, it can still be observed with waves (including millimeter waves) that propagate through or around walls. It can be used to perform an evaluation of their potential applications in medical uses, and apart from life-saving operations, it can also be used in the military to see enemies behind walls. The application of the device is aimed at detecting the locations of individuals trapped under debris in catastrophic events such as earthquakes, with the objective being to maintain a stance of objective benevolence.

Removable Coating for In-Space Cold Welding Applications

The present invention is a removable coating to enable cold welding in space applications. The base metal aluminum is prepared with a coating that prevents excessive oxidation of the metal within Earth’s atmosphere. Once in space, this coating is designed to be easily removable in order to reveal clean metal that can be pressed together to form strong, hermetic seals for manufacturing large structures in space.