26 August 2019 in IEEE Transactions on Visualization and Computer Graphics
Combining data content with visual embellishments, infographics can effectively deliver messages in an engaging and memorable manner. Various authoring tools have been proposed to facilitate the creation of infographics. However, creating a professional infographic with these authoring tools is still not an easy task, requiring much time and design expertise. Therefore, these tools are generally not attractive to casual users, who are either unwilling to take time to learn the tools or lacking in proper design expertise to create a professional infographic. In this paper, we explore an alternative approach: to automatically generate infographics from natural language statements. We first conducted a preliminary study to explore the design space of infographics. Based on the preliminary study, we built a proof-of-concept system that automatically converts statements about simple proportionrelated statistics to a set of infographics with pre-designed styles. Finally, we demonstrated the usability and usefulness of the system through sample results, exhibits, and expert reviews.
28 April 2023 in Nature Communications
Retrosynthesis analysis is an important task in organic chemistry with numerous industrial applications. Previously, machine learning approaches employing natural language processing techniques achieved promising results in this task by first representing reactant molecules as strings and subsequently predicting reactant molecules using text generation or machine translation models. Chemists cannot readily derive useful insights from traditional approaches that rely largely on atom-level decoding in the string representations, because human experts tend to interpret reactions by analyzing substructures that comprise a molecule. It is well-established that some substructures are stable and remain unchanged in reactions. In this paper, we developed a substructure-level decoding model, where commonly preserved portions of product molecules were automatically extracted with a fully data-driven approach. Our model achieves improvement over previously reported models, and we demonstrate that its performance can be boosted further by enhancing the accuracy of these substructures. Analyzing substructures extracted from our machine learning model can provide human experts with additional insights to assist decision-making in retrosynthesis analysis.
2 February 2024 in Digital Discovery
Retro-BLEU is a statistical metric to evaluate the plausibility of model-generated retrosynthesis routes based on reaction template sequences analysis.
8 June 2023 in Journal of Cheminformatics
Retrosynthesis is an important task in organic chemistry. Recently, numerous data-driven approaches have achieved promising results in this task. However, in practice, these data-driven methods might lead to sub-optimal outcomes by making predictions based on the training data distribution, a phenomenon we refer as frequency bias. For example, in template-based approaches, low-ranked predictions are typically generated by less common templates with low confidence scores which might be too low to be comparable, and it is observed that recorded reactants can be among these low-ranked predictions. In this work, we introduce RetroRanker, a ranking model built upon graph neural networks, designed to mitigate the frequency bias in predictions of existing retrosynthesis models through re-ranking. RetroRanker incorporates potential reaction changes of each set of predicted reactants in obtaining the given product to lower the rank of chemically unreasonable predictions. The predicted re-ranked results on publicly available retrosynthesis benchmarks demonstrate that we can achieve improvement on most state-of-the-art models with RetroRanker. Our preliminary studies also indicate that RetroRanker can enhance the performance of multi-step retrosynthesis.
1 October 2011 in International Journal of Applied Evolutionary Computation
Rule-based Genetic Algorithms (GAs) have been used in the application of pattern classification (Corcoran & Sen, 1994), but conventional GAs have weaknesses. First, the time spent on learning is long. Moreover, the classification accuracy achieved by a GA is not satisfactory. These drawbacks are due to existing undesirable features embedded in conventional GAs. The number of rules within the chromosome of a GA classifier is usually set and fixed before training and is not problem-dependent. Secondly, conventional approaches train the data in batch without considering whether decomposition solves the problem. Thirdly, when facing large-scale real-world problems, GAs cannot utilise resources efficiently, leading to premature convergence. Based on these observations, this paper develops a novel algorithmic framework that features automatic domain and task decomposition and problem-dependent chromosome length (rule number) selection to resolve these undesirable features. The proposed Recursive Learning of Genetic Algorithm with Task Decomposition and Varied Rule Set (RLGA) method is recursive and trains and evolves a team of learners using the concept of local fitness to decompose the original problem into sub-problems. RLGA performs better than GAs and other related solutions regarding training duration and generalization accuracy according to the experimental results.