2023-24 Takeda Fellows: Advancing analysis on the intersection of AI and well being | MIT Information


The College of Engineering has chosen 13 new Takeda Fellows for the 2023-24 tutorial yr. With help from Takeda, the graduate college students will conduct pathbreaking analysis starting from distant well being monitoring for digital scientific trials to ingestible units for at-home, long-term diagnostics.

Now in its fourth yr, the MIT-Takeda Program, a collaboration between MIT’s College of Engineering and Takeda, fuels the event and utility of synthetic intelligence capabilities to learn human well being and drug growth. A part of the Abdul Latif Jameel Clinic for Machine Studying in Well being, this system coalesces disparate disciplines, merges idea and sensible implementation, combines algorithm and {hardware} improvements, and creates multidimensional collaborations between academia and business.

The 2023-24 Takeda Fellows are:

Adam Gierlach

Adam Gierlach is a PhD candidate within the Division of Electrical Engineering and Pc Science. Gierlach’s work combines progressive biotechnology with machine studying to create ingestible units for superior diagnostics and supply of therapeutics. In his earlier work, Gierlach developed a non-invasive, ingestible machine for long-term gastric recordings in free-moving sufferers. With the help of a Takeda Fellowship, he’ll construct on this pathbreaking work by creating sensible, energy-efficient, ingestible units powered by application-specific built-in circuits for at-home, long-term diagnostics. These revolutionary units — able to figuring out, characterizing, and even correcting gastrointestinal illnesses — symbolize the forefront of biotechnology. Gierlach’s progressive contributions will assist to advance basic analysis on the enteric nervous system and assist develop a greater understanding of gut-brain axis dysfunctions in Parkinson’s illness, autism spectrum dysfunction, and different prevalent issues and circumstances.

Vivek Gopalakrishnan

Vivek Gopalakrishnan is a PhD candidate within the Harvard-MIT Program in Well being Sciences and Know-how. Gopalakrishnan’s purpose is to develop biomedical machine-learning strategies to enhance the examine and therapy of human illness. Particularly, he employs computational modeling to advance new approaches for minimally invasive, image-guided neurosurgery, providing a secure different to open mind and spinal procedures. With the help of a Takeda Fellowship, Gopalakrishnan will develop real-time pc imaginative and prescient algorithms that ship high-quality, 3D intraoperative picture steering by extracting and fusing info from multimodal neuroimaging information. These algorithms may permit surgeons to reconstruct 3D neurovasculature from X-ray angiography, thereby enhancing the precision of machine deployment and enabling extra correct localization of wholesome versus pathologic anatomy.

Hao He

Hao He’s a PhD candidate within the Division of Electrical Engineering and Pc Science. His analysis pursuits lie on the intersection of generative AI, machine studying, and their functions in drugs and human well being, with a selected emphasis on passive, steady, distant well being monitoring to help digital scientific trials and health-care administration. Extra particularly, He goals to develop reliable AI fashions that promote equitable entry and ship honest efficiency unbiased of race, gender, and age. In his previous work, He has developed monitoring techniques utilized in scientific research of Parkinson’s illness, Alzheimer’s illness, and epilepsy. Supported by a Takeda Fellowship, He’ll develop a novel expertise for the passive monitoring of sleep levels (utilizing radio signaling) that seeks to deal with present gaps in efficiency throughout totally different demographic teams. His undertaking will deal with the issue of imbalance in out there datasets and account for intrinsic variations throughout subpopulations, utilizing generative AI and multi-modality/multi-domain studying, with the purpose of studying sturdy options which can be invariant to totally different subpopulations. He’s work holds nice promise for delivering superior, equitable health-care companies to all individuals and will considerably impression well being care and AI.

Chengyi Lengthy

Chengyi Lengthy is a PhD candidate within the Division of Civil and Environmental Engineering. Lengthy’s interdisciplinary analysis integrates the methodology of physics, arithmetic, and pc science to research questions in ecology. Particularly, Lengthy is creating a collection of probably groundbreaking strategies to elucidate and predict the temporal dynamics of ecological techniques, together with human microbiota, that are important topics in well being and medical analysis. His present work, supported by a Takeda Fellowship, is concentrated on creating a conceptual, mathematical, and sensible framework to know the interaction between exterior perturbations and inner neighborhood dynamics in microbial techniques, which can function a key step towards discovering bio options to well being administration. A broader perspective of his analysis is to develop AI-assisted platforms to anticipate the altering conduct of microbial techniques, which can assist to distinguish between wholesome and unhealthy hosts and design probiotics for the prevention and mitigation of pathogen infections. By creating novel strategies to deal with these points, Lengthy’s analysis has the potential to supply highly effective contributions to drugs and world well being.

Omar Mohd

Omar Mohd is a PhD candidate within the Division of Electrical Engineering and Pc Science. Mohd’s analysis is concentrated on creating new applied sciences for the spatial profiling of microRNAs, with doubtlessly essential functions in most cancers analysis. By way of progressive combos of micro-technologies and AI-enabled picture evaluation to measure the spatial variations of microRNAs inside tissue samples, Mohd hopes to realize new insights into drug resistance in most cancers. This work, supported by a Takeda Fellowship, falls inside the rising subject of spatial transcriptomics, which seeks to know most cancers and different illnesses by inspecting the relative areas of cells and their contents inside tissues. The final word purpose of Mohd’s present undertaking is to search out multidimensional patterns in tissues that will have prognostic worth for most cancers sufferers. One worthwhile part of his work is an open-source AI program developed with collaborators at Beth Israel Deaconess Medical Middle and Harvard Medical College to auto-detect most cancers epithelial cells from different cell varieties in a tissue pattern and to correlate their abundance with the spatial variations of microRNAs. By way of his analysis, Mohd is making progressive contributions on the interface of microsystem expertise, AI-based picture evaluation, and most cancers therapy, which may considerably impression drugs and human well being.

Sanghyun Park

Sanghyun Park is a PhD candidate within the Division of Mechanical Engineering. Park specializes within the integration of AI and biomedical engineering to deal with complicated challenges in human well being. Drawing on his experience in polymer physics, drug supply, and rheology, his analysis focuses on the pioneering subject of in-situ forming implants (ISFIs) for drug supply. Supported by a Takeda Fellowship, Park is at present creating an injectable formulation designed for long-term drug supply. The first purpose of his analysis is to unravel the compaction mechanism of drug particles in ISFI formulations by complete modeling and in-vitro characterization research using superior AI instruments. He goals to realize an intensive understanding of this distinctive compaction mechanism and apply it to drug microcrystals to realize properties optimum for long-term drug supply. Past these basic research, Park’s analysis additionally focuses on translating this data into sensible functions in a scientific setting by animal research particularly aimed toward extending drug launch length and bettering mechanical properties. The progressive use of AI in creating superior drug supply techniques, coupled with Park’s worthwhile insights into the compaction mechanism, may contribute to bettering long-term drug supply. This work has the potential to pave the way in which for efficient administration of continual illnesses, benefiting sufferers, clinicians, and the pharmaceutical business.

Huaiyao Peng

Huaiyao Peng is a PhD candidate within the Division of Organic Engineering. Peng’s analysis pursuits are targeted on engineered tissue, microfabrication platforms, most cancers metastasis, and the tumor microenvironment. Particularly, she is advancing novel AI strategies for the event of pre-cancer organoid fashions of high-grade serous ovarian most cancers (HGSOC), an particularly deadly and difficult-to-treat most cancers, with the purpose of gaining new insights into development and efficient remedies. Peng’s undertaking, supported by a Takeda Fellowship, can be one of many first to make use of cells from serous tubal intraepithelial carcinoma lesions discovered within the fallopian tubes of many HGSOC sufferers. By inspecting the mobile and molecular modifications that happen in response to therapy with small molecule inhibitors, she hopes to establish potential biomarkers and promising therapeutic targets for HGSOC, together with personalised therapy choices for HGSOC sufferers, in the end bettering their scientific outcomes. Peng’s work has the potential to result in essential advances in most cancers therapy and spur progressive new functions of AI in well being care. 

Priyanka Raghavan

Priyanka Raghavan is a PhD candidate within the Division of Chemical Engineering. Raghavan’s analysis pursuits lie on the frontier of predictive chemistry, integrating computational and experimental approaches to construct highly effective new predictive instruments for societally essential functions, together with drug discovery. Particularly, Raghavan is creating novel fashions to foretell small-molecule substrate reactivity and compatibility in regimes the place little information is accessible (essentially the most life like regimes). A Takeda Fellowship will allow Raghavan to push the boundaries of her analysis, making progressive use of low-data and multi-task machine studying approaches, artificial chemistry, and robotic laboratory automation, with the purpose of making an autonomous, closed-loop system for the invention of high-yielding natural small molecules within the context of underexplored reactions. Raghavan’s work goals to establish new, versatile reactions to broaden a chemist’s artificial toolbox with novel scaffolds and substrates that might kind the idea of important medication. Her work has the potential for far-reaching impacts in early-stage, small-molecule discovery and will assist make the prolonged drug-discovery course of considerably sooner and cheaper.

Zhiye Music

Zhiye “Zoey” Music is a PhD candidate within the Division of Electrical Engineering and Pc Science. Music’s analysis integrates cutting-edge approaches in machine studying (ML) and {hardware} optimization to create next-generation, wearable medical units. Particularly, Music is creating novel approaches for the energy-efficient implementation of ML computation in low-power medical units, together with a wearable ultrasound “patch” that captures and processes photos for real-time decision-making capabilities. Her current work, performed in collaboration with clinicians, has centered on bladder quantity monitoring; different potential functions embrace blood stress monitoring, muscle prognosis, and neuromodulation. With the help of a Takeda Fellowship, Music will construct on that promising work and pursue key enhancements to present wearable machine applied sciences, together with creating low-compute and low-memory ML algorithms and low-power chips to allow ML on sensible wearable units. The applied sciences rising from Music’s analysis may supply thrilling new capabilities in well being care, enabling highly effective and cost-effective point-of-care diagnostics and increasing particular person entry to autonomous and steady medical monitoring.

Peiqi Wang

Peiqi Wang is a PhD candidate within the Division of Electrical Engineering and Pc Science. Wang’s analysis goals to develop machine studying strategies for studying and interpretation from medical photos and related scientific information to help scientific decision-making. He’s creating a multimodal illustration studying method that aligns data captured in massive quantities of medical picture and textual content information to switch this data to new duties and functions. Supported by a Takeda Fellowship, Wang will advance this promising line of labor to construct sturdy instruments that interpret photos, study from sparse human suggestions, and motive like medical doctors, with doubtlessly main advantages to essential stakeholders in well being care.

Oscar Wu

Haoyang “Oscar” Wu is a PhD candidate within the Division of Chemical Engineering. Wu’s analysis integrates quantum chemistry and deep studying strategies to speed up the method of small-molecule screening within the growth of recent medication. By figuring out and automating dependable strategies for locating transition state geometries and calculating barrier heights for brand spanking new reactions, Wu’s work may make it attainable to conduct the high-throughput ab initio calculations of response charges wanted to display screen the reactivity of huge numbers of energetic pharmaceutical elements (APIs). A Takeda Fellowship will help his present undertaking to: (1) develop open-source software program for high-throughput quantum chemistry calculations, specializing in the reactivity of drug-like molecules, and (2) develop deep studying fashions that may quantitatively predict the oxidative stability of APIs. The instruments and insights ensuing from Wu’s analysis may assist to rework and speed up the drug-discovery course of, providing important advantages to the pharmaceutical and medical fields and to sufferers.

Soojung Yang

Soojung Yang is a PhD candidate within the Division of Supplies Science and Engineering. Yang’s analysis applies cutting-edge strategies in geometric deep studying and generative modeling, together with atomistic simulations, to raised perceive and mannequin protein dynamics. Particularly, Yang is creating novel instruments in generative AI to discover protein conformational landscapes that provide higher velocity and element than physics-based simulations at a considerably decrease price. With the help of a Takeda Fellowship, she is going to construct upon her profitable work on the reverse transformation of coarse-grained proteins to the all-atom decision, aiming to construct machine-learning fashions that bridge a number of measurement scales of protein conformation range (all-atom, residue-level, and domain-level). Yang’s analysis holds the potential to supply a robust and broadly relevant new device for researchers who search to know the complicated protein features at work in human illnesses and to design medication to deal with and treatment these illnesses.

Yuzhe Yang

Yuzhe Yang is a PhD candidate within the Division of Electrical Engineering and Pc Science. Yang’s analysis pursuits lie on the intersection of machine studying and well being care. In his previous and present work, Yang has developed and utilized progressive machine-learning fashions that tackle key challenges in illness prognosis and monitoring. His many notable achievements embrace the creation of one of many first machine learning-based options utilizing nocturnal respiration alerts to detect Parkinson’s illness (PD), estimate illness severity, and observe PD development. With the help of a Takeda Fellowship, Yang will develop this promising work to develop an AI-based prognosis mannequin for Alzheimer’s illness (AD) utilizing sleep-breathing information that’s considerably extra dependable, versatile, and economical than present diagnostic instruments. This passive, in-home, contactless monitoring system — resembling a easy residence Wi-Fi router — can even allow distant illness evaluation and steady development monitoring. Yang’s groundbreaking work has the potential to advance the prognosis and therapy of prevalent illnesses like PD and AD, and it presents thrilling prospects for addressing many well being challenges with dependable, inexpensive machine-learning instruments.