The College of Engineering has chosen 13 new Takeda Fellows for the 2023-24 educational 12 months. With help from Takeda, the graduate college students will conduct pathbreaking analysis starting from distant well being monitoring for digital medical trials to ingestible units for at-home, long-term diagnostics.

Now in its fourth 12 months, the MIT-Takeda Program, a collaboration between MIT’s College of Engineering and Takeda, fuels the event and software of synthetic intelligence capabilities to learn human well being and drug improvement. 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 revolutionary 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 good, 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 ailments — signify the forefront of biotechnology. Gierlach’s revolutionary 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 Expertise. Gopalakrishnan’s purpose is to develop biomedical machine-learning strategies to enhance the examine and remedy of human illness. Particularly, he employs computational modeling to advance new approaches for minimally invasive, image-guided neurosurgery, providing a protected various to open mind and spinal procedures. With the help of a Takeda Fellowship, Gopalakrishnan will develop real-time laptop imaginative and prescient algorithms that ship high-quality, 3D intraoperative picture steerage by extracting and fusing info from multimodal neuroimaging knowledge. These algorithms might enable 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 purposes in medication and human well being, with a specific emphasis on passive, steady, distant well being monitoring to help digital medical trials and health-care administration. Extra particularly, He goals to develop reliable AI fashions that promote equitable entry and ship honest efficiency impartial of race, gender, and age. In his previous work, He has developed monitoring methods utilized in medical research of Parkinson’s illness, Alzheimer’s illness, and epilepsy. Supported by a Takeda Fellowship, He’ll develop a novel know-how for the passive monitoring of sleep levels (utilizing radio signaling) that seeks to handle current gaps in efficiency throughout totally different demographic teams. His venture will sort out the issue of imbalance in accessible datasets and account for intrinsic variations throughout subpopulations, utilizing generative AI and multi-modality/multi-domain studying, with the purpose of studying strong options which can be invariant to totally different subpopulations. He’s work holds nice promise for delivering superior, equitable health-care providers to all folks and will considerably influence 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 laptop science to analyze questions in ecology. Particularly, Lengthy is creating a collection of probably groundbreaking methods to elucidate and predict the temporal dynamics of ecological methods, 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 methods, 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 habits of microbial methods, 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 handle these points, Lengthy’s analysis has the potential to supply highly effective contributions to medication 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 purposes in most cancers analysis. By revolutionary 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 ailments by inspecting the relative places of cells and their contents inside tissues. The last word purpose of Mohd’s present venture is to search out multidimensional patterns in tissues which will have prognostic worth for most cancers sufferers. One beneficial 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 sorts in a tissue pattern and to correlate their abundance with the spatial variations of microRNAs. By his analysis, Mohd is making revolutionary contributions on the interface of microsystem know-how, AI-based picture evaluation, and most cancers remedy, which might considerably influence medication 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 handle advanced 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 the moment 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 via complete modeling and in-vitro characterization research using superior AI instruments. He goals to realize a radical 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 information into sensible purposes in a medical setting via animal research particularly geared toward extending drug launch period and bettering mechanical properties. The revolutionary use of AI in creating superior drug supply methods, coupled with Park’s beneficial insights into the compaction mechanism, might contribute to bettering long-term drug supply. This work has the potential to pave the way in which for efficient administration of power ailments, 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 centered on engineered tissue, microfabrication platforms, most cancers metastasis, and the tumor microenvironment. Particularly, she is advancing novel AI methods 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 venture, supported by a Takeda Fellowship, will likely 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 remedy with small molecule inhibitors, she hopes to establish potential biomarkers and promising therapeutic targets for HGSOC, together with personalised remedy choices for HGSOC sufferers, finally bettering their medical outcomes. Peng’s work has the potential to result in essential advances in most cancers remedy and spur revolutionary new purposes 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 purposes, together with drug discovery. Particularly, Raghavan is creating novel fashions to foretell small-molecule substrate reactivity and compatibility in regimes the place little knowledge is out there (probably the most reasonable regimes). A Takeda Fellowship will allow Raghavan to push the boundaries of her analysis, making revolutionary 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 medicine. 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 Track

Zhiye “Zoey” Track is a PhD candidate within the Division of Electrical Engineering and Pc Science. Track’s analysis integrates cutting-edge approaches in machine studying (ML) and {hardware} optimization to create next-generation, wearable medical units. Particularly, Track 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 photographs for real-time decision-making capabilities. Her latest work, performed in collaboration with clinicians, has centered on bladder quantity monitoring; different potential purposes embody blood stress monitoring, muscle analysis, and neuromodulation. With the help of a Takeda Fellowship, Track will construct on that promising work and pursue key enhancements to current wearable machine applied sciences, together with creating low-compute and low-memory ML algorithms and low-power chips to allow ML on good wearable units. The applied sciences rising from Track’s analysis might provide 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 photographs and related medical knowledge to help medical decision-making. He’s creating a multimodal illustration studying strategy that aligns information captured in giant quantities of medical picture and textual content knowledge to switch this information to new duties and purposes. Supported by a Takeda Fellowship, Wang will advance this promising line of labor to construct strong instruments that interpret photographs, study from sparse human suggestions, and purpose 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 improvement of latest medicine. By figuring out and automating dependable strategies for locating transition state geometries and calculating barrier heights for brand spanking new reactions, Wu’s work might make it potential to conduct the high-throughput ab initio calculations of response charges wanted to display the reactivity of huge numbers of lively pharmaceutical substances (APIs). A Takeda Fellowship will help his present venture 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 might assist to rework and speed up the drug-discovery course of, providing vital 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 higher perceive and mannequin protein dynamics. Particularly, Yang is creating novel instruments in generative AI to discover protein conformational landscapes that supply better velocity and element than physics-based simulations at a considerably decrease value. With the help of a Takeda Fellowship, she’s 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 variety (all-atom, residue-level, and domain-level). Yang’s analysis holds the potential to offer a robust and broadly relevant new device for researchers who search to know the advanced protein capabilities at work in human ailments and to design medicine to deal with and treatment these ailments.

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 revolutionary machine-learning fashions that handle key challenges in illness analysis and monitoring. His many notable achievements embody the creation of one of many first machine learning-based options utilizing nocturnal respiration indicators to detect Parkinson’s illness (PD), estimate illness severity, and observe PD development. With the help of a Takeda Fellowship, Yang will broaden this promising work to develop an AI-based analysis mannequin for Alzheimer’s illness (AD) utilizing sleep-breathing knowledge that’s considerably extra dependable, versatile, and economical than present diagnostic instruments. This passive, in-home, contactless monitoring system — resembling a easy house Wi-Fi router — may even allow distant illness evaluation and steady development monitoring. Yang’s groundbreaking work has the potential to advance the analysis and remedy of prevalent ailments like PD and AD, and it presents thrilling potentialities for addressing many well being challenges with dependable, inexpensive machine-learning instruments.