Browse Articles

Although there has been great progress in the field of Natural language processing (NLP) over the last few years, particularly with the development of attention-based models, less research has contributed towards modeling keystroke log data. State of the art methods handle textual data directly and while this has produced excellent results, the time complexity and resource usage are quite high for such methods. Additionally, these methods fail to incorporate the actual writing process when assessing text and instead solely focus on the content. Therefore, we proposed a framework for modeling textual data using keystroke-based features. Such methods pay attention to how a document or response was written, rather than the final text that was produced. These features are vastly different from the kind of features extracted from raw text but reveal information that is otherwise hidden. We hypothesized that pairing efficient machine learning techniques with keystroke log information should produce results comparable to transformer techniques, models which pay more or less attention to the different components of a text sequence in a far quicker time. Transformer-based methods dominate the field of NLP currently due to the strong understanding they display of natural language. We showed that models trained on keystroke log data are capable of effectively evaluating the quality of writing and do it in a significantly shorter amount of time compared to traditional methods. This is significant as it provides a necessary fast and cheap alternative to increasingly larger and slower LLMs.

Reinforcement learning (RL) is a form of machine learning that can be harnessed to develop artificial intelligence by exposing the intelligence to multiple generations of data. The study demonstrates how reply buffer reward mechanics can inform the creation of new pruning methods to improve RL efficiency.

With molecular energy being an integral element to the study of molecules and molecular interactions, computational methods to determine molecular energy are used for the preservation of time and resources. However, these computational methods have high demand for computer resources, limiting their widespread feasibility. The authors of this study employed machine learning to address this disadvantage, utilizing neural networks trained on different representations of molecules to predict molecular properties without the requirement of computationally-intensive processing. In their findings, the authors determined the Feedforward Neural Network, trained by two separate models, as capable of predicting molecular energy with limited prediction error.

Every year, around 40% of undergraduate students in the United States discontinue their studies, resulting in a loss of valuable education for students and a loss of money for colleges. Even so, colleges across the nation struggle to discover the underlying causes of these high dropout rates. In this paper, the authors discuss the use of machine learning to find correlations between the built environment factors and the retention rates of colleges. They hypothesized that one way for colleges to improve their retention rates could be to improve the physical characteristics of their campus to be more pleasing. The authors used image classification techniques to look at images of colleges and correlate certain features like colors, cars, and people to higher or lower retention rates. With three possible options of high, medium, and low retention rates, the probability that their models reached the right conclusion if they simply chose randomly was 33%. After finding that this 33%, or 0.33 mark, always fell outside of the 99% confidence intervals built around their models’ accuracies, the authors concluded that their machine learning techniques can be used to find correlations between certain environmental factors and retention rates.

The authors develop and test a machine learning algorithm for predicting diabetes diagnoses.

Using a convolution neural network, these authors show machine learning can clinically diagnose breast cancer with high accuracy.

Zoos offer educational and scientific advantages but face high maintenance costs and challenges in animal care due to diverse species' habits. Challenges include tracking animals, detecting illnesses, and creating suitable habitats. We developed a deep learning framework called SmartZoo to address these issues and enable efficient animal monitoring, condition alerts, and data aggregation. We discovered that the data generated by our model is closer to real data than random data, and we were able to demonstrate that the model excels at generating data that resembles real-world data.

Coronary heart disease (CHD) is the leading cause of death in the U.S., responsible for nearly 700,000 deaths in 2021, and is marked by artery clogging that can lead to heart attacks. Traditional prediction methods require expensive clinical tests, but a new study explores using machine learning on demographic, clinical, and behavioral survey data to predict CHD.

The challenge of accurately identifying shark species is crucial for biodiversity monitoring but is often hindered by time-consuming and labor-intensive manual methods. To address this, SharkNet, a CNN model based on AlexNet, achieved 93% accuracy in classifying shark species using a limited dataset of 1,400 images across 14 species. SharkNet offers a more efficient and reliable solution for marine biologists and conservationists in species identification and environmental monitoring.

People with Type One diabetes often rely on Continuous Blood Glucose Monitors (CGMs) to track their blood glucose and manage their condition. Researchers are now working to help people with Type One diabetes more easily monitor their health by developing models that will future blood glucose levels based on CGM readings. Jalla and Ghanta tackle this issue by exploring the use of AI models to forecast blood glucose levels with CGM data.