Current research projects

  • The Big Data to Knowledge (BD2K) Centers Coordination Center

    The Big Data to Knowledge (BD2K) Centers Coordination Center

    Cri­tical to the success of the NIH BD2K Ini­tia­tive is a specia­lized entity for pro­mo­ting com­mon inte­rests and ensu­ring sus­tai­ned impact of inno­va­tions birt­hed by the mul­tiple BD2K Cen­ters of Excel­lence (COEs). Both the imme­diate and long-­term success of the BD2K Ini­tia­tive is depen­dent upon how deve­lo­ped resources and tools are sha­red, inte­gra­ted and mobi­lized, as well as effec­ti­vely adop­ted by the broad scien­ti­fic com­mu­ni­ty. UCLA is home to the BD2K Cen­ters-Coordinating Cen­ter (CCC). The ...

    More

  • The Center for Domain-Specific Computing

    The Center for Domain-Specific Computing

    The cur­rent Cen­ter for Domain-S­peci­fic Com­pu­ting (CDSC) is a col­la­bo­ra­tion between UCLA, Rice Uni­ver­si­ty, Ore­gon Health Sciences Uni­ver­sity (OH­SU), and Intel Research. The scope of its efforts inclu­des leve­ra­ging the research results from our past work, but sig­ni­ficantly expan­ding and exten­ding the research along the fol­lowing research areas: 1. Acce­le­ra­tor-cent­ric arc­hi­tec­tu­res (ACAs) in which novel met­ho­do­lo­gies and algo­rithms to auto­ma­tically ext­ract acce­le­ra­tor buil­ding blocks (ABBs) i...

    More

  • An Observational Stroke Model for Decision Support

    An Observational Stroke Model for Decision Support

    Annual­ly, it is esti­ma­ted that more than 795,000 Ame­ricans expe­rience a stro­ke. The seve­rity of neu­ro­lo­gical damage due to an acute stroke is miti­ga­ted by the early res­to­ra­tion of blood flow to the affec­ted area; and more people are now sur­vi­ving stro­kes through ear­lier inter­ven­tion with throm­bo­ly­tic agents and inter­ven­tional clot ret­rie­val devices. Unfor­tu­na­te­ly, the rapid deve­lop­ment of new drugs and devices in this area has made it dif­ficult to pro­vide treat­ment gui­dance for a given pat...

    More

  • A Predictive Prognostic Model for Brain Cancer

    A Predictive Prognostic Model for Brain Cancer

    Each year, almost half of all diag­no­sed pri­mary brain tumors in the Uni­ted Sta­tes are Grade IV glioblas­toma mul­ti­forme (GBMs). While recent efforts have begun to unco­ver the gene­tic pathways invol­ved in this cancer's etio­logy – and poten­tial met­hods for treat­ment – arguably, no speci­fic prog­nos­tic model has ari­sen (and been suf­ficiently vali­da­ted) to pro­vide widespread usa­bi­lity and indi­vi­dually tai­lo­red pre­dic­tions about a patient's prog­no­sis, let alone sug­gest optim...

    More

  • The Los Angeles PRISMS Center

    The Los Angeles PRISMS Center

    Sup­por­ted by NIH/­NI­BIB U54 EB022002 (PI: Alex Bui)

    The Los Ange­les Pediat­ric Research Inte­gra­ting Sen­sor Moni­to­ring Sys­tems (LA PRISMS) Cen­ter is fos­te­ring the deve­lop­ment and applica­tion of mobile health (mHealth) tech­no­lo­gies that dee­pen our scien­ti­fic unders­tan­ding and cli­nical mana­ge­ment of pediat­ric con­di­tions. Brin­ging toget­her lea­ding experts from UCLA and USC in bio­me­dical infor­ma­tics, com­pu­ter science, wire­less health, envi­ron­men­tal science and health, and pediat­rics, t...

    More

  • Research Maps for Integrating Research and Planning Experiments

    Research Maps for Integrating Research and Planning Experiments

    Causa­lity is a cent­ral concept for both basic science and cli­nical medici­ne. In the last few deca­des, we have seen sig­ni­ficant deve­lop­ment of mat­he­ma­tical for­ma­lisms for mode­ling causa­li­ty. Des­pite the exis­tence of robust and expres­sive for­ma­lisms for causal mode­ling, such for­ma­lisms are surpri­singly unde­rused by bio­lo­gists see­king to iden­tify causal mec­ha­nisms and by cli­nicians see­king to unders­tand the etio­logy of disea­se. MII is thus wor­king to adapt sta­te-of-t­he-art causal-­disco­very me...

    More

  • RUMI: Retrieving Understandable Medical Information

    RUMI: Retrieving Understandable Medical Information

    While deci­sion-­ma­king for any cancer is complex, this is especially true for lung cancer as its treat­ment options often involve mode­rate to sig­ni­ficant mor­bi­di­ties, and the disease itself has a high mor­ta­lity rate. A patient's abi­lity to cope with cancer is influenced by a num­ber of fac­tors, inclu­ding know­ledge about the disease and the resul­tant sense of empower­ment to unders­tand and par­tici­pate in his/­her own healthcare and to make deci­sions. Inc­rea­singly, onco­logy patients are goin...

    More

  • Topic Models for Automatic Summarization of Patient Records

    Topic Models for Automatic Summarization of Patient Records

    Pri­mary care phy­sicians (PCPs) are res­pon­sible for reviewing and unders­tan­ding a wide spect­rum of a patient's medical his­tory in order to make infor­med deci­sions regar­ding care. Howe­ver, a variety of fac­tors impede this process, inclu­ding: the inc­rea­sing complexity and num­ber of diag­nos­tic tests and treat­ments, health infor­ma­tion exc­hange stan­dards that may add more infor­ma­tion to the medical record, and the need to efficiently see more patients in less time. The use of topic models f...

    More