- would like any opportunities to do sbc on diiferential equation structure, stratified structure - evaluating and improving posterior inference for difficult posterior - could you walk me each project and your ideas? - preference - hierarchical, dynamic http://localhost:63029/stanify.html - epidemiology or public health - applied methodologist motivated, language, concepts, good match (ok for you first choice), varies in different universities (funding - project funding, work the goal) aalto - specific research plan, Fcai (requirement is topic should be btw two professors) sbc- doubtful ablout martin's idea; useful - stay commonly used (computationally heavy, easy) new tools might be able to use data, helping martin (minimize computation time, criterion -- if the estimate is really fast, rank 0 or 1), we don't want hypothesis testing try to estimate amount of bias (sbc would reveal - infer is not perfect, rhat with many chain) how important variations of cross validation, proj pred, variable selection (model diagnostic ideas), 50% faster - enjoy research, flexiibility in deciding when and what topics to work, - administration, thesis (not great), too much work (gradual chg, ), doctoral studnt need paper (publish stress, arxiv only), productive enough (andrew - visibility, famous, ppl noticed, ) - professors having life in u.s. (harping text) I need a serious advice on whether I, before-phd student 1. Topic: **Bayesian workflows for iterative model building and networks of models** We formalize and develop theory and diagnostics for iterative Bayesian model building. The practical workflow recommendations and diagnostics guide the modeller through the appropriate steps to ensure safe iterative model building, or indicate - jukka pekka onnela (likelihood free inference, simulation-based), approximate bayesian compution - likelihood free inference = sbc (highly ) - depending on doing recently your background (we've only discussion, my octorla stdent only partial, ) - us-based professor (bayescomp program - which professors) - mentioned sbc coauthor when the modeler is likely to be in the danger zone. 2. Topic: **Evaluating and improving posterior inference for difficult posteriors** Both MCMC and distributional approximations often struggle to handle complex posteriors, but we lack good tools for understanding how and why. We study diagnostics for identifying the nature of the computational difficulty, e.g. whether the difficulty is caused by narrow funnels or strong curvature. We also develop improved inference algorithms, e.g. via automated and semi-automated transformations. 3. Topic: **Workflows for better priors** Bayesian models rely on prior distributions that encode knowledge about the problem, but specifying good priors is often difficult in practice. We are working on multiple fronts on making it easier, with contributions to e.g. prior elicitation, prior diagnostics, prior checking, and specification of priors in predictive spaces. hierarchical spline and gaussian process for low risk/high what I would do ## simulation-based calibration The work I have done with Dr. Hazhir Rahmandad at Sloan is, hierarchical Bayesian model calibration to measure vaccine value. Different states in U.S. is calibrated separately and I believe this technique is applicable to mammogram data for different age groups or hospitals. Apart from that, I would be excited if I can find an application for simulation-based calibration, which I developed for the past two years. This is a readily-implemented procedure that can identify sources of poorly implemented analyses, including biased computational algorithms or incorrect model specifications as detailed in [this](https://streaklinks.com/BXUMzsYLyzyj9kcsrQTtdkNQ/http%3A%2F%2Fwww.stat.columbia.edu%2F~gelman%2Fresearch%2Funpublished%2Fsbc.pdf?email=hyunji.moonb%40gmail.com) paper. leaving out the prior knowledge on structure from t ## how to embed the prior knowledge on structure - without this, variance would explode simulation-based operations science - the need for a small model, would - - [Application of the Cross-Entropy Method to the Buffer Allocation Problem in a Simulation-Based Environment](https://people.smp.uq.edu.au/DirkKroese/ps/aorbap.pdf) -