Machine Learning for Biomedical Data Science

随着科学不断寻求改善对健康状况的分析和治疗的方法，生物信息学-生物学和计算机科学的结合-已经成为一个相对较新的领域，在这个领域中，收集、处理和应用数据来解决生物问题，在分子水平上解开了谜团。生物信息学寻求在数据集中发现相关的“隐藏”信息，以生成解决生物数据科学当前挑战所需的解决方案。处于生物信息学研究前沿的李一峰博士在人工智能(AI)和机器学习方面的工作过滤和识别了关键的遗传和物理信息，这些信息目前要么隐藏在过多的数据中，要么由于数据太少而等待发现。李博士研究项目的主要目标是开发新的基于算法的机器学习方法：1.教会人工智能更好地解释复杂的生物数据/生物医学图像分析；2.通过优化和强化学习技术提高解决生物技术中多目标设计问题的效率；3.通过使用迁移学习技术来填补由于生物数据不足而造成的空白；以及4.通过迁移学习技术来提高解决生物技术中多目标设计问题的效率。可以用于现实生活中的应用，从而推动数据科学领域跨越各种学科。从非处方药头痛药物到抗癌药物，每种治疗方法都必须测试其达到体内指定目标的能力并具有预期的效果。李博士的研究探索了过滤海量数据的新方法，开发了一种算法，通过字面意思“教”软件更好地理解数据集中的生物信息，然后实现所有必要的目标，目的是改进和简化药物设计。在数据太少的情况下--比如X光、核磁共振和其他需要足够的图像才能成为有效诊断技术的医疗技术--李博士和他的团队正在开发新的算法，以填补成像方面的空白，从而改进生物医学图像分析。李的研究项目专注于创造新的计算机模型和方法，以应对各种现实生活应用中的数据科学挑战。他的工作使新兴的生物信息学领域能够最大限度地提高其有效性，并最终提高人工智能支持所有加拿大人的健康和福祉的能力。

As science continually seeks ways to improve the analysis and remedy of health conditions, bioinformatics - the union of biology and computer science - has emerged as a relatively new field in which the collection, processing, and application of data to biological questions unlock mysteries at the molecular level. Bioinformatics seeks to discover within data sets the relevant, "hidden" information needed to generate solutions that address current challenges in biological data science.Situated at the forefront of bioinformatics research, Dr. Yifeng Li's work in artificial intelligence (AI) and machine learning filters and identifies crucial genetic and physical information that is currently either buried in too much data or waiting to be discovered due to too little data. The key objective of Dr. Li's research program is to develop new algorithm-based machine learning approaches that:1. teach AI to better interpret complex biological data/biomedical image analysis;2. increase efficiencies in solving multi-objective design problems in biotechnology through optimization and reinforcement learning techniques;3. fill gaps created by insufficient biological data by using transfer learning techniques; and4. can be used for real-life applications and thereby advance the data science field across a variety of disciplines.From over-the-counter headache medications to cancer drugs, each treatment must be tested for its ability to reach the designated target within the body and have the desired effect. Dr. Li's research explores new methods of filtering through massive amounts of data by developing algorithms that literally "teach" software to better understand the biological information within a data set and then fulfill all required objectives, with the goal of improving and streamlining drug design. In cases with too little data - such as in x-rays, MRIs, and other medical technologies that require enough images to be an effective diagnostic technique - Dr. Li and his team are developing new algorithms to fill gaps in imaging and thereby improve biomedical image analysis.Dr. Li's research program focuses on creating new computer models and methods that address data science challenges in a variety of real-life applications. His work allows the emergent field of bioinformatics to maximize its effectiveness and ultimately advance the capacity of AI to support the health and well-being of all Canadians.