A5065777975- Machine Learning and ELM
- Domain Adaptation and Few-Shot Learning
- Topic Modeling
- Natural Language Processing Techniques
- Speech Recognition and Synthesis
- Stochastic Gradient Optimization Techniques
- Remote-Sensing Image Classification
- Online Learning and Analytics
- Advanced Text Analysis Techniques
- Face and Expression Recognition
- Artificial Intelligence in Healthcare and Education
- Spectroscopy Techniques in Biomedical and Chemical Research
- Privacy-Preserving Technologies in Data
- Machine Learning in Healthcare
- Face recognition and analysis
- Time Series Analysis and Forecasting
- Teaching and Learning Programming
- FinTech, Crowdfunding, Digital Finance
- Extracellular vesicles in disease
- Text and Document Classification Technologies
- Career Development and Diversity
- Neural Networks and Applications
- Experimental Learning in Engineering
- Stock Market Forecasting Methods
- Scientific Computing and Data Management
Lawrence Technological University
2019-2024
Hampton University
2024
Oakland University
2016-2018
Spatiotemporal fusion is important in providing high spatial resolution earth observations with a dense time series, and recently, learning-based methods have been attracting broad interest. These algorithms project image patches onto feature space the enforcement of simple mapping to predict fine from corresponding coarse ones. However, sophisticated projection, e.g., sparse representation, always computationally complex difficult be implemented on large patches, which cannot grasp enough...
A course-based undergraduate research experience (CURE) involves the incorporation of at under-graduate level, typically as a replacement to lab portion course. motivational driver for incorporating in classroom is ensure greater participation and retention students science, technology, engineering, mathematics (STEM). Studies have also shown that who engage experiences gain better understanding scientific process are more likely pursue graduate studies. The CURE model regarded pedagogical...
Word embeddings are a low-dimensional vector representation of words that incorporates context. TWo popular methods word2vec and global vectors (GloVe). Word2vec is single-hidden layer feedforward neural network (SLFN) has an auto-encoder influence for computing word context matrix using backpropagation training. GloVe computes the first then performs factorization on to arrive at embeddings. Backpropagation typical training method SLFN's, which time consuming requires iterative tuning....
The graphical user interface has been the predominate in web and software applications with interaction primarily being click-driven. Advances Natural Language Understanding (NLU) are leading transition to interfaces that conversation-driven using natural language. Utilizing a state-of-the-art open source framework NLU capabilities, we demonstrate illustrate conversational for Stock Analysis based on real-time computation processing textual numerical data. key features of proposed system...
Generative AI assistants are AI-powered applications that can provide personalized responses to user queries or prompts. A variety of have recently been released, and among the most popular is OpenAI's ChatGPT. In this work-in-progress in innovative practice, we explore evidence-based learning strategies integration for computer science engineering education. We expect research will lead pedagogical approaches enhance undergraduate particular, describe how ChatGPT was used two...
Clinical texts are inherently complex due to the medical domain expertise required for content comprehension. In addition, unstructured nature of these narratives poses a challenge automatically extracting information. natural language processing, use word embeddings an effective approach generate representations (vectors) in low dimensional space. this paper we log-linear model (a type neural model) and Linear Discriminant Analysis with kernel-based Extreme Learning Machine (ELM) map...
Stacked or sequential convolutional layers in a Convolutional Neural Network (CNN) have shown state-of-the-art results Image and Pattern Recognition. Recently, CNN's promising Natural Language Processing (NLP) tasks. Using CNN with concurrent layers, we conduct text categorization on clinical narrative dataset imbalance classes. Clinical narratives are written natural language, documenting the encounter as observed from clinician along process of care. For this research, experiment various...
Introduction We describe herein a large-scale, multidisciplinary course-based undergraduate research experience program (CRE) developed at Lawrence Technological University (LTU). In our program, all students enrolled in CRE classes participate authentic experiences within the framework of curriculum, eliminating self-selection processes and other barriers to traditional extracurricular experiences. Methods Since 2014, we have designed implemented more than 40 courses College Arts Sciences...
Artificial Intelligence (AI) has grown dramatically over the past few decades and much greater influence today on human performance in every walk of life. AI is great for accelerating our work, but there are also downsides to it. One such downside application Facial Recognition Technologies (FRT) available today, which adverse consequences like racial discrimination or wrongful judgement by commercial firms even law enforcement organizations. The landmark 'Gender Shades' project [1] 2018...
Weighted word co-occurrence frequencies are considered the bedrock of embeddings. Also known as a low-dimensional numerical representation, embeddings capture pair extracted from corpus in an unsupervised manner. The rendering can be two-step process with first s tep i nvolving t he building context matrix then using factorization method to reduce dimensionality. In this research study, constructed scratch and Truncated Singular Value Decomposition is applied matrix. Five experimental values...
Subword embeddings are integral to transformers such as Generative Pre-trained Transformer (GPT) and Bidirectional Encoder Representations from Transformers (BERT), which utilized in various natural language processing tasks. A subword is a subset of word can be decomposed into one or more subwords characters. One challenge with tokenization determining the optimal tokens for representing word. This research proposes refinement GPT tokenizer based on an analysis GPT. The proposed approach...