Current drug discovery processes can cost billions of dollars and usually take five to ten years. People have been researching and implementing various computational approaches to search for molecules and compounds from the chemical space, which can be on the order of 1060 molecules. One solution involves deep generative models, which are artificial intelligence models that learn from nonlinear data by modeling the probability distribution of chemical structures and creating similar data points from the trends it identifies. Aiming for faster runtime and greater robustness when analyzing high-dimensional data, we designed and implemented a Hybrid Quantum-Classical Generative Adversarial Network (QGAN) to synthesize molecules.
Recurrent neural networks (RNNs) are useful for text generation since they can generate outputs in the context of previous ones. Baroque music and language are similar, as every word or note exists in context with others, and they both follow strict rules. The authors hypothesized that if we represent music in a text format, an RNN designed to generate language could train on it and create music structurally similar to Bach’s. They found that the music generated by our RNN shared a similar structure with Bach’s music in the input dataset, while Bachbot’s outputs are significantly different from this experiment’s outputs and thus are less similar to Bach’s repertoire compared to our algorithm.
The authors test the feasibility of using thermoelectric modules as a power source and as an air conditioner to decrease reliance on fossil fuels. The results showed that, at its peak, their battery generated 27% more power – in watts per square inch – than a solar panel, and the thermoelectric air conditioner operated despite an unsteady input voltage. The battery has incredible potential, especially if its peak power output can be maintained.
A key barrier to adoption of solar energy technology is the low efficiency of solar cells converting solar energy into electricity. Sims and Sims tackle this problem by coding a Raspberry Pi as a multimeter to determine which wavelength of light generates the most voltage and current from a solar panel.
In this study, the authors studied the ability for bacteria to develop antibiotic resistance over successive generations and modeled the trajectory to predict how antibiotic resistance is developed.
Here, the authors sought to identify a method to optimize the lift generated by an airfoil based solely on its shape. By beginning with a Bernoullian model to predict an optimized wing shape, the authors then tested their model against other possible shapes by constructing them from Styrofoam and testing them in a small wind tunnel. Contrary to their hypothesis, they found their expected optimal airfoil shape did not result in the greatest lift generation. They attributed this to a variety of confounding variables and concluded that their results pointed to a correlation between airfoil shape and lift generation.
The application of machine learning techniques has facilitated the automatic annotation of behavior in video sequences, offering a promising approach for ethological studies by reducing the manual effort required for annotating each video frame. Nevertheless, before solely relying on machine-generated annotations, it is essential to evaluate the accuracy of these annotations to ensure their reliability and applicability. While it is conventionally accepted that there cannot be a perfect annotation, the degree of error associated with machine-generated annotations should be commensurate with the error between different human annotators. We hypothesized that machine learning supervised with adequate human annotations would be able to accurately predict body parts from video sequences. Here, we conducted a comparative analysis of the quality of annotations generated by humans and machines for the body parts of sheep during treadmill walking. For human annotation, two annotators manually labeled six body parts of sheep in 300 frames. To generate machine annotations, we employed the state-of-the-art pose-estimating library, DeepLabCut, which was trained using the frames annotated by human annotators. As expected, the human annotations demonstrated high consistency between annotators. Notably, the machine learning algorithm also generated accurate predictions, with errors comparable to those between humans. We also observed that abnormal annotations with a high error could be revised by introducing Kalman Filtering, which interpolates the trajectory of body parts over the time series, enhancing robustness. Our results suggest that conventional transfer learning methods can generate behavior annotations as accurate as those made by humans, presenting great potential for further research.