Search for: All records

Predicting chemical reaction yields is pivotal for efficient chemical synthesis, an area that focuses on the creation of novel compounds for diverse uses. Yield prediction demands accurate representations of reactions for forecasting practical transformation rates. Yet, the uncertainty issues broadcasting in real-world situations prohibit current models to excel in this task owing to the high sensitivity of yield activities and the uncertainty in yield measurements. Existing models often utilize single-modal feature representations, such as molecular fingerprints, SMILES sequences, or molecular graphs, which is not sufficient to capture the complex interactions and dynamic behavior of molecules in reactions. In this paper, we present an advanced Uncertainty-Aware Multimodal model (UAM) to tackle these challenges. Our approach seamlessly integrates data sources from multiple modalities by encompassing sequence representations, molecular graphs, and expert-defined chemical reaction features for a comprehensive representation of reactions. Additionally, we address both the model and data-based uncertainty, refining the model’s predictive capability. Extensive experiments on three datasets, including two high throughput experiment (HTE) datasets and one chemist-constructed Amide coupling reaction dataset, demonstrate that UAM outperforms the stateof-the-art methods. The code and used datasets are available at https://github.com/jychen229/Multimodal-reaction-yieldprediction. 

Inference of joule-class THz radiation sources from microchannel targets driven with hundreds of joule, picosecond lasers is reported. THz sources of this magnitude are useful for nonlinear pumping of matter and for charged-particle acceleration and manipulation. Microchannel targets demonstrate increased laser–THz conversion efficiency compared to planar foil targets, with laser energy to THz energy conversion up to ∼0.9% in the best cases.

 

Coastal marsh survival may be compromised by sea‐level rise, limited sediment supply, and subsidence. Storms represent a fundamental forcing for sediment accumulation in starving marshes because they resuspend bottom material in channels and tidal flats and transport it to the marsh surface. However, it is unrealistic to simulate at high resolution all storms that occurred in the past decades to obtain reliable sediment accumulation rates. Similarly, it is difficult to cover all possible combinations of water levels and wind conditions in fictional scenarios. Thus, we developed a new method that derives long‐term deposition rates from short‐term deposition generated by a finite number of storms. Twelve storms with different intensity and frequency were selected in Terrebonne Bay, Louisiana, USA and simulated with the 2D Delft3D‐FLOW model coupled with the Simulating Waves Nearshore (SWAN) module. Storm impact was analyzed in terms of geomorphic work, namely the product of deposition and frequency. To derive the long‐term inorganic mass accumulation rates, the new method generates every possible combination of the 12 chosen storms and uses a linear model to fit modeled inorganic deposition with measured inorganic mass accumulation rates. The linear model with the best fit (highestR²) was used to derive a map of inorganic mass accumulation rates. Results show that a storm with 1.7 ± 1.6 years return period provides the largest geomorphic work, suggesting that the most impactful storms are those that balance intensity with frequency. Model results show higher accumulation rates in marshes facing open areas where waves can develop and resuspend sediments. This method has the advantage of considering only a few real scenarios and can be applied in any marsh‐bay system.

 

This paper considers the problem of error in variables identification for switched affine models. Since it is well known that this problem is generically NP hard, several relaxations have been proposed in the literature. However, while these approaches work well for low dimensional systems with few subsystems, they scale poorly with both the number of subsystems and their memory. To address this difficulty, we propose a computationally efficient alternative, based on embedding the data in the manifold of positive semidefinite matrices, and using a manifold metric there to perform the identification. Our main result shows that, under dwell-time assumptions, the proposed algorithm is convergent, in the sense that it is guaranteed to identify the system for suitably low noise. In scenarios with larger noise levels, we provide experimental results showing that the proposed method outperforms existing ones. The paper concludes by illustrating these results with academic examples and a non-trivial application: action video segmentation. 

At San Francisco State University, a Hispanic Serving Institute and a Primarily Undergraduate Institution, 67% of engineering students are from ethnic minority groups, with only 27% of Hispanic students retained and graduated in their senior year. Additionally, only 14% of students reported full-time employment secured at the time of graduation. Of these secured jobs, only 54% were full-time positions (40+ hours a week). To improve the situation, San Francisco State University, in collaboration with two local community colleges, Skyline and Cañada Colleges, was recently funded by the National Science Foundation through a Hispanic Serving Institute Improving Undergraduate STEM Education Strengthening Student Motivation and Resilience through Research and Advising program to enhance undergraduate engineering education and build capacity for student success. This project will use a data-driven and evidence-based approach to identify the barriers to the success of underrepresented minority students and to generate new knowledge on the best practices for increasing students’ retention and graduation rates, self- efficacy, professional development, and workforce preparedness. Three objectives underpin this overall goal. The first is to develop and implement a Summer Research Internship Program together with community college partners. The second is to establish an HSI Engineering Success Center to provide students with academic resources, networking opportunities with industry, and career development tools. The third is to develop resources for the professional development of faculty members, including Summer Faculty Teaching Workshops, an Inclusive Teaching and Mentoring Seminar Series, and an Engineering Faculty Learning Community. Qualitative and quantitative approaches are used to assess the project outcomes using a survey instrument and interview protocols developed by an external evaluator. This paper discusses an overview of the project and its first-year implementation. The focus is placed on the introduction and implementation of the several main project components, namely the Engineering Success Center, Summer Research Internship Program, and Faculty Summer Teaching Workshop. The preliminary evaluation results, demonstrating the great success of these strategies, are also discussed.