KAIX 석학 강연

KAIX 석학강연

The purpose of this talk is to mathematically investigate the formation of a plasma sheath, and to analyze the Bohm criterions which are required for the formation. Bohm derived originally the (hydrodynamic) Bohm criterion from the Euler–Poisson system. Boyd and Thompson proposed the (kinetic) Bohm criterion from kinetic point of view, and then Riemann derived it from the Vlasov–Poisson system. We study the solvability of boundary value problems of the Vlasov–Poisson system. On the process, we see that the kinetic Bohm criterion is a necessary condition for the solvability. The argument gives a simpler derivation of the criterion. Furthermore, the hydrodynamic criterion can be derived from the kinetic criterion. It is of great interest to find the relation between the solutions of the Vlasov–Poisson and Euler–Poisson systems. To clarify the relation, we also investigate the hydrodynamic limit of solutions of the Vlasov–Poisson system.

The nonorientable four-ball genus of a knot $K$ in $S^3$ is the minimal first Betti number of nonorientable surfaces in $B^4$ bounded by $K$. By amalgamating ideas from involutive knot Floer homology and unoriented knot Floer homology, we give a new lower bound on the smooth nonorientable four-ball genus $\gamma_4$ of any knot. This bound is sharp for several families of torus knots, including $T_{4n,(2n\pm 1)^2}$ for even $n\ge2$, a family Longo showed were counterexamples to Batson's conjecture. We also prove that, whenever $p$ is an even positive integer and $\frac{p}{2}$ is not a perfect square, the torus knot $T_{p,q}$ does not bound a locally flat M{\" o}bius band for almost all integers $q$ relatively prime to $p$.

Anyons are quasiparticles in two dimensions. They do not belong to the two classes of elementary particles, bosons and fermions. Instead, they obey Abelian or non-Abelian fractional statistics. Their quantum mechanical states are determined by fusion or braiding, to which braid groups and conformal field theories are naturally applied. Some of non-Abelian anyons are central in realization of topological qubits and topological quantum computing. I will introduce the basic properties of anyons and their recent experimental signatures observed in systems of topological order such as fractional quantum Hall systems and topological superconductors.

The cohomology of local Shimura varieties, and of more general spaces of local shtukas, is of fundamental interest in the Langlands program. On the one hand, it is supposed to realize instances of the local Langlands correspondence. On the other hand, there is a tight relationship with the cohomology of global Shimura varieties. In recent joint work with Kaletha and Weinstein, we proved the first general results towards the Kottwitz conjecture, which predicts how supercuspidal L-packets contribute to the cohomology of local shtuka spaces. I will review this whole story, and give some overview of the ideas which enter into our proof. The key idea in our argument - namely, that the Kottwitz conjecture should follow from some form of the Lefschetz-Verdier fixed point formula - was already formulated by Michael Harris in the '90s. However, executing this idea brings substantial technical challenges. I will try to emphasize the new ingredients which allow us to implement this idea in full generality.

---

Please contact Wansu Kim at for Zoom meeting info and any inquiry. For the list of Number Theory seminar talks, please visit the KAIST Number Theory seminar webpage. https://sites.google.com/site/wansukimmaths/kants-kaist-number-theory-seminar

Chronotherapeutics- that is administering drugs following the patient’s biological rhythms over the 24 h span- may largely impact on both drug toxicities and efficacy in various pathologies including cancer [1]. However, recent findings highlight the critical need of personalizing circadian delivery according to the patient sex, genetic background or chronotype. Chronotherapy personalization requires to reliably account for the temporal dynamics of molecular pathways of patient’s response to drug administration [2]. In a context where clinical molecular data is usually minimal in individual patients, multi-scale- from preclinical to clinical- systems pharmacology stands as an adapted solution to describe gene and protein networks driving circadian rhythms of treatment efficacy and side effects and allow for the design of personalized chronotherapies. Such a multiscale approach is being undertaken for personalizing the circadian administration of irinotecan, one of the cornerstones of chemotherapies against digestive cancers. Irinotecan molecular chronopharmacology was studied at the cellular level in an in vitro/in silico investigation. Large transcription rhythms of period T= 28 h 06 min (SD 1 h 41 min) moderated drug bioactivation, detoxification, transport, and target in synchronized Caco-2 colorectal cancer cell cultures. These molecular rhythms translated into statistically significant changes according to drug timing in irinotecan pharmacokinetics, pharmacodynamics, and drug-induced apoptosis. Clock silencing through siBMAL1 exposure ablated all the chronopharmacology mechanisms. Mathematical modeling highlighted circadian bioactivation and detoxification as the most critical determinants of irinotecan chronopharmacology [3]. The cellular model of irinotecan chronoPK-PD was further tested on SW480 and SW620 cell lines, and connected to a new clock model to investigate the feasibility of irinotecan timing personalization solely based on clock gene expression monitoring (Hesse, Martinelli et al., under review). To step towards the clinics, on one side, mathematical models of irinotecan, oxaliplatin and 5-fluorouracil pharmacokinetics were designed to precisely compute the exposure concentration of tissue over time after complex chronomodulated drug administration through programmable pumps [4]. On the other side, we aimed to design a model learning methodology predicting from non-invasively measured circadian biomarkers (e.g. rest-activity, body temperature, cortisol, food intake, melatonin), the patient peripheral circadian clocks and associated optimal drug timing [5]. We investigated at the molecular scale the influence of systemic regulators on peripheral clocks in four classes of mice (2 strains, 2 sexes). Best models involved a modulation of either Bmal1 or Per2 transcription most likely by temperature or nutrient exposure cycles. The strengths of systemic regulations were found to be significantly different according to mouse sex and genetic background. References 1. Ballesta, A., et al., Systems Chronotherapeutics. Pharmacol Rev, 2017. 69(2): p. 161-199. 2. Sancar, A. and R.N. Van Gelder, Clocks, cancer, and chronochemotherapy. Science, 2021. 371(6524). 3. Dulong, S., et al., Identification of Circadian Determinants of Cancer Chronotherapy through In Vitro Chronopharmacology and Mathematical Modeling. Mol Cancer Ther, 2015. 4. Hill, R.J.W., et al., Optimizing circadian drug infusion schedules towards personalized cancer chronotherapy. PLoS Comput Biol, 2020. 16(1): p. e1007218. 5. Martinelli, J., et al., Model learning to identify systemic regulators of the peripheral circadian clock. 2021.

---

This talk will be presented online. Zoom link: 709 120 4849 (pw: 1234)

Bouchet (1987) defined delta-matroids by relaxing the base exchange axiom of matroids. Oum (2009) introduced a graphic delta-matroid from a pair of a graph and its vertex subset. We define a $\Gamma$-graphic delta-matroid for an abelian group $\Gamma$, which generalizes a graphic delta-matroid. For an abelian group $\Gamma$, a $\Gamma$-labelled graph is a graph whose vertices are labelled by elements of $\Gamma$. We prove that a certain collection of edge sets of a $\Gamma$-labelled graph forms a delta-matroid, which we call a $\Gamma$-graphic delta-matroid, and provide a polynomial-time algorithm to solve the separation problem, which allows us to apply the symmetric greedy algorithm of Bouchet (1987) to find a maximum weight feasible set in such a delta-matroid. We also prove that a $\Gamma$-graphic delta-matroid is a graphic delta-matroid if and only if it is even. We prove that every $\mathbb{Z}_p^k$-graphic delta matroid is represented by some symmetric matrix over a field of characteristic of order $p^k$, and if every $\Gamma$-graphic delta-matroid is representable over a finite field $\mathbb{F}$, then $\Gamma$ is isomorphic to $\mathbb{Z}_p^k$ and $\mathbb{F}$ is a field of order $p^\ell$ for some prime $p$ and positive integers $k$ and $\ell$. This is joint work with Duksang Lee and Sang-il Oum.

In the first part, I introduce a novel variational model for the joint enhancement and restoration of dark images corrupted by blurring and/or noise. The model decomposes a given dark image into reflectance and illumination images that are recovered from blurring and/or noise. In addition, our approach utilizes non-convex total variation regularization on all variables. This allows us to adequately denoise homogeneous regions while preserving the details and edges in both reflectance and illumination images, which leads to clean and sharp final enhanced images. Experimental results demonstrate the effectiveness of the proposed model when compared to other state-of-the-art methods in terms of both visual aspect and image quality measures. In the second part, I propose a novel variational model for the restoration of a single color image degenerated by haze. The model extends the total variation based model, by inserting an inter-channel correlation term. This additional term permits both color and gray-valued transmission maps, which enable broader applications of the proposed model. Numerical experiments validate the outstanding performance of the proposed model compared to the state-of-the-art methods.

Within a given species, fluctuations in egg or embryo size is unavoidable. Despite this, the gene expression pattern and hence the embryonic structure often scale in proportion with the body length. This scaling phenomenon is very common in development and regeneration and has long fascinated scientists. I will first discuss a generic theoretical framework to show how scaling gene expression pattern can emerge from non-scaling morphogen gradients. I will then demonstrate that the Drosophila gap gene system achieves scaling in a way that is entirely consistent with our theory. Remarkably, a parameter-free model based on the theory quantitatively accounts for the gap gene expression pattern in nearly all morphogen mutants. Furthermore, the regulation logic and the coding/decoding strategy of the gap gene system can be revealed. Our work provides a general theoretical framework on a large class of problems where scaling output is induced by non-scaling input, as well as a unified understanding of scaling, mutants’ behavior and regulation in the Drosophila gap gene and related systems.

---

This talk will be presented online. Zoom link: 709 120 4849 (pw: 1234)

The next few talks will be more like learning than research: I will explain some preparation material, which is considered "well-known" by the experts, but which I didn't find a reference for in the form I need. My next goal is to explain the proof that the Picard group of the so-called quotient of a torsor of a simply connected simple split algebraic group modulo a Borel subgroup does not change under field extension. In the first talk I will explain the basic machinery to prove this fact, namely Galois descent theory. Given a variety X over a non-algebraically closed field F with no or "not enough" rational points, Galois descent theory allows one to work with an extension K of F and with X_K and study the properties of the original X. If there is enough time, I will also define torsors and show how to construct them using Galois descent.

Partial differential equations such as heat equations have traditionally been our main tool to study physical systems. However, physical systems are affected by randomness (noise). Thus, stochastic partial differential equations have gained popularity as an alternative. In this talk, we first consider what “noise” means mathematically and then consider stochastic heat equations perturbed by space-time white noise such as parabolic Anderson model and stochastic reaction-diffusion equations (e.g., KPP or Allen-Cahn equations). Those stochastic heat equations have similar properties as heat equations, but exhibit different behavior such as intermittency and dissipation, especially as time increases. We investigate in this talk how the long-time behaviors of the stochastic heat equations are different from heat equations.

In this talk, we are going to discuss boundary regularities of various degenerate local equation and nonlocal equations. Diffusion rates deform undefined geometry related to diffusion and the corresponding distance function makes important role in the theory of regularity. And then we will also discuss the possible applications.

The temporal credit assignment, the problem of determining which actions in the past are responsible for the current outcome (long-term cause and effect), is difficult to solve because one needs to backpropagate the error signal through space and time. Despite its computational challenges, humans are very good at solving this problem. Our lab uses reinforcement learning theory and algorithms to explore the nature of computations underlying the brain’s ability to solve the temporal credit assignment. I will outline two-fold approaches to this issue: (1) training a computational model from human behavioral data without underfitting and overfitting (Brain → AI) and (2) using the trained model to manipulate the way the human brain solves the temporal credit assignment problem (AI → brain).

---

Education/employments PhD, KAIST (2009)Postdoc, MIT (2010-2011), Caltech (2011-2015)Faculty, KAIST (2015-now) Honors/awards IBM Academic Research Award (2021)Google Faculty Research Award (2017)Della-Martin Fellowship (2014) KAIST Breakthroughs (2020)KAIST Songam Distinguished Research Award (2019)KAIST Top 10 Technologies (2019)KAIST Institute Faculty Award (2019) KIIS Young Investigator Award (2016)ICROS Young Investigator Award (2016)

Given a sequence of random i.i.d. 2 by 2 complex matrices, it is a classical problem to study the statistical properties of their product. This theory dates back to fundamental works of Furstenberg, Kesten, etc. and is still an active research topic. In this talk, I intend to show how methods from complex analysis and analogies with holomorphic dynamics offer a new point of view to this problem. This is used to obtain several new limit theorems for these random processes, often in their optimal version. This is based on joint works with T.-C. Dinh and H. Wu.

Majority dynamics on a graph $G$ is a deterministic process such that every vertex updates its $\pm 1$-assignment according to the majority assignment on its neighbor simultaneously at each step. Benjamini, Chan, O'Donnell, Tamuz and Tan conjectured that, in the Erd\H{o}s--R\'enyi random graph $G(n,p)$, the random initial $\pm 1$-assignment converges to a $99\%$-agreement with high probability whenever $p=\omega(1/n)$. This conjecture was first confirmed for $p\geq\lambda n^{-1/2}$ for a large constant $\lambda$ by Fountoulakis, Kang and Makai. Although this result has been reproved recently by Tran and Vu and by Berkowitz and Devlin, it was unknown whether the conjecture holds for $p< \lambda n^{-1/2}$. We break this $\Omega(n^{-1/2})$-barrier by proving the conjecture for sparser random graphs $G(n,p)$, where $\lambda' n^{-3/5}\log n \leq p \leq \lambda n^{-1/2}$ with a large constant $\lambda'>0$.

We continue our discussion on the result of Marden, Thurston and Bonahon which states that in hyperbolic 3-manifolds, every immersed surface of which the fundamental group is invectively embedded in the 3-manifold group is quasi-fuchsian or doubly degenerated. Surface subgroups of 3-manifold groups play an important rule in 3-manifold theory. For instance, some collection of immersed surfaces give rise to a CAT(0) cube complex. Especially, in the usual construction of the CAT(0) cube complex, each immersed surface composing the collection is quasi-fuchsian. In this talk, I introduce the work by Cooper, Long and Reid. In hyperbolic mapping tori, the work gives a criterion to determine whether the given immersed surface is quasi-fuchsian or not. The criterion is given in terms of laminations induced in immersed surfaces.

Immune sentinel cells must initiate the appropriate immune response upon sensing the presence of diverse pathogens or immune stimuli. To generate stimulus-specific gene expression responses, immune sentinel cells have evolved a temporal code in the dynamics of stimulus responsive transcription factors. I will present recent works 1) using an information theoretic approach to identify the codewords, termed “signaling codons”, 2) using a machine learning approach to characterize their reliability and points of confusion, and 3) dynamical systems modeling to characterize the molecular circuits that allow for their encoding. I will present progress on how the temporal code may be decoded to specify immune responses. Further, I will discuss to what extent such a code may be harnessed to achieve greater pharmacological specificity when therapeutically targeting pleiotropic signaling hubs. NFκB Signaling: information theory, signaling codons

---

This talk will be presented online. Zoom link: 709 120 4849 (pw: 1234)

For a quadratic projective variety X ⊂P_r , the locus of quadratic polynomials of rank 3 in the homogeneous ideal I(X) defines a projective algebraic set, say PHI_3(X), in P(I, (X)_2). So, it provides several projective invariants of X. In this talk, I will speak about the structure of Phi_2(C) when C ⊂P_n is a rational normal curve. This is based on the joint work with Saerom Shim.

In his famous 1900 presentation, Hilbert proposed so-called the Hilbert’s 6thproblem, namely “Mathematical Treatment of the Axioms of Physics”. He mentioned that “Boltzmann's work on the principles of mechanics suggests the problem of developing mathematically the limiting processes, there merely indicated, which lead from the atomistic view to the laws of motion of continua.” In this lecture, we present some recent development of the Hilbert’s 6th problem in the Boltzmann theory when the various fluid models have natural “singularities” such as unbounded vorticity and formation of boundary layers.

이번 발표를 통해 수리과학 모델을 이용한 감염병 확산 예측 방법 그리고 방역 정책의 감염 확산 억제 효과 분석에 대하여 소개하겠습니다. 감염 확산 모형의 가장 기본이면서 널리 쓰이고 있는 compartment model, 그리고 지역 단위 인구 이동 자료를 반영한 metapopulation model에 대하여 논의하고, 시시각각 변화하는 감염 확산 상황을 표현하기 적합한 data assimilation method을 살펴보겠습니다. 방역 정책 효과 분석을 위한 수리과학 모델로서 microsimulation model을 소개하겠습니다. Microsimulation model은 정부의 정책 변화가 사회, 경제적으로 미치는 영향을 분석하고자 제안된 시뮬레이션 도구로 거시적 수준의 경제, 사회, 인구 변화를 각 개인과 가구 단위의 미시적 사건들로부터 기술합니다. Microsimulation model을 이용하면 가구, 직장/학교, 종교 및 친목 모임의 밀접 접촉을 통한 호흡기 감염병 확산을 시뮬레이션할 수 있습니다. 그리고 휴교령, 직장 재택 근무, 종교 시설 폐쇄 등의 비약물적 조치가 감염병 확산 방지에 어떤 효과를 지니는지 분석할 수 있다는 장점이 있습니다.

---

https://us02web.zoom.us/j/82312487069?pwd=RUJFUmVaZnBYdzJNOUZ5TTRIbzJXZz09

The cohomology of Shimura varieties have rich structures and have been studied for many years. Some new vanishing theorems were proved in the last few years and especially the one by Caraiani-Scholze is crucial in arithmetic applications. I will survey these results, and discuss further development.

---

(Please contact Wansu Kim at for Zoom meeting info and any inquiry.)

In my first talk I am going to speak about Schubert calculus. Let G/B be a flag variety, where G is a linear simple algebraic group, and B is a Borel subgroup. Schubert calculus studies (in classical terms) multiplication in the cohomology ring of a flag variety over the complex numbers, or (in more algebraic terms) the Chow ring of the flag variety. This ring is generated as a group by the classes of so-called Schubert varieties (or their Poincare duals, if we speak about the classical cohomology ring), i. e. of the varieties of the form BwB/B, where w is an element of the Weyl group. As a ring, it is almost generated by the classes of Schubert varieties of codimension 1, called Schubert divisors. More precisely, the subring generated by Schubert divisors is a subgroup of finite index. These two facts lead to the following general question: how to decompose a product of Schubert divisors into a linear combination of Schubert varieties. In my talk, I am going to address (and answer if I have time) two more particular versions of this question: If G is of type A, D, or E, when does a coefficient in such a linear combination equal 0? When does it equal 1?

I am going to present an algorithm for computing a feedback vertex set of a unit disk graph of size k, if it exists, which runs in time $2^{O(\sqrt{k})}(n + m)$, where $n$ and $m$ denote the numbers of vertices and edges, respectively. This improves the $2^{O(\sqrt{k}\log k)}(n + m)$-time algorithm for this problem on unit disk graphs by Fomin et al. [ICALP 2017].

Recently, deep learning approaches have become the main research frontier for image reconstruction and enhancement problems thanks to their high performance, along with their ultra-fast inference times. However, due to the difficulty of obtaining matched reference data for supervised learning, there has been increasing interest in unsupervised learning approaches that do not need paired reference data. In particular, self-supervised learning and generative models have been successfully used for various inverse problem applications. In this talk, we overview these approaches from a coherent perspective in the context of classical inverse problems and discuss their various applications. In particular, the cycleGAN approach and a recent Noise2Score approach for unsupervised learning will be explained in detail using optimal transport theory and Tweedie’s formula with score matching.

This is joint work with Kenjiro Ishizuka (Kyoto). We study global behavior of solutions to the nonlinear Klein-Gordon equation with a damping and a focusing nonlinearity on the Euclidean space. Recently, Cote, Martel and Yuan proved the soliton resolution conjecture completely in the one-dimensional case: every global solution in the energy space is asymptotic to a superposition of solitons getting away from each other as time tends to infinity. The next question is to see which initial data evolve into each of the asymptotic forms. The asymptotic decomposition is very sensitive to initial perturbation because all the solitons are unstable. We consider the simplest non-trivial setting in general space dimensions: the global behavior of solutions starting near a superposition of two ground states. Cote, Martel, Yuan and Zhao proved that the solutions asymptotic to 2-solitons form a codimension-2 manifold in the energy space. Our question is what happens for the other initial data in the neighborhood. As an answer, we give a complete classification of those solutions into 5 types of global behavior. Two of them are asymptotic to the positive ground state and the negative one respectively. They form two codimension-1 manifolds that are joined at their boundary by the Cote-Martel-Yuan-Zhao manifold of 2-solitons. The connected union of those three manifolds separates the remainder of the neighborhood into the open set of global decaying solutions and that of blow-up. The main difficulty to prove it is in controlling the direction of instability in two dimensions attached to the two soliton components, because the soliton interactions are not integrable in time, breaking the simple superposition of the linearized approximation around each soliton. It is resolved by showing that the non-integrable interactions do not essentially affect the direction of instability, using the reflection symmetry of the equation and the 2-solitons. I will also explain the difficulty for the 3-solitons due to a more dramatic phenomenon, which may be called soliton merger.

(KAIX Distinguished Lectures Series)

---

1st talk 10:00-11:00 short break and Q&A 11:00-11:15 2nd talk 11:15-12:15 Q&A 12:15-12:30

The extremal function $c(H)$ of a graph $H$ is the supremum of densities of graphs not containing $H$ as a minor, where the density of a graph is the ratio of the number of edges to the number of vertices. Myers and Thomason (2005), Norin, Reed, Thomason and Wood (2020), and Thomason and Wales (2019) determined the asymptotic behaviour of $c(H)$ for all polynomially dense graphs $H$, as well as almost all graphs of constant density. We explore the asymptotic behavior of the extremal function in the regime not covered by the above results, where in addition to having constant density the graph $H$ is in a graph class admitting strongly sublinear separators. We establish asymptotically tight bounds in many cases. For example, we prove that for every planar graph $H$, \[c(H) = (1+o(1))\max (v(H)/2, v(H)-\alpha(H)),\] extending recent results of Haslegrave, Kim and Liu (2020). Joint work with Sergey Norin and David R. Wood.

---

https://youtube.com/c/ibsdimag

I will explain how to put certain natural geometric structures on Tate-Shafarevich groups and other related groups attached to abelian varieties over function fields. We can refine arithmetic duality theorems by taking these geometric structures into account. This has applications to Weil-etale cohomology, the Birch-Swinnerton-Dyer conjecture and Iwasawa theory. Partially based on joint work with Geisser and with Lai, Longhi, Tan and Trihan.

---

Please contact Wansu Kim at for Zoom meeting info and any inquiry.

In hyperbolic 3 manifolds, by Marden, Thurston and Bonahon, every immersed surface of which the fundamental group is invectively embedded in the 3-manifold group is quasi-fuchsian or doubly degenerated. Surface subgroups of 3-manifold groups play an important rule in 3-manifold theory. For instance, some collection of immersed surfaces give a rise to a CAT(0) cube complex. Especially, in the usual construction of the CAT(0) cube complex, each immersed surface composing the collection is quasi-fuchsian. In this talk, I introduce the work by Cooper, Long and Reid. In hyperbolic mapping tori, the work gives a criterion to determine whether the given immersed surface is quasi-fuchsian or not. The criterion is given in terms of laminations induced in immersed surfaces.

Many of real-world data, e.g., the VGGFace2 dataset, which is a collection of multiple portraits of individuals, come with nested structures due to grouped observation. The Ornstein auto-encoder (OAE) is an emerging framework for representation learning from nested data, based on an optimal transport distance between random processes. An attractive feature of OAE is its ability to generate new variations nested within an observational unit, whether or not the unit is known to the model. A previously proposed algorithm for OAE, termed the random-intercept OAE (RIOAE), showed an impressive performance in learning nested representations, yet lacks theoretical justification. In this work, we show that RIOAE minimizes a loose upper bound of the employed optimal transport distance. After identifying several issues with RIOAE, we present the product-space OAE (PSOAE) that minimizes a tighter upper bound of the distance and achieves orthogonality in the representation space. PSOAE alleviates the instability of RIOAE and provides more flexible representation of nested data. We demonstrate the high performance of PSOAE in the three key tasks of generative models: exemplar generation, style transfer, and new concept generation. This is a joint work with Dr. Youngwon Choi (UCLA) and Sungdong Lee (SNU).

Questions about the mechanistic operation of biological systems are naturally formulated as stochastic processes, but confronting such models with data can be challenging. In this talk, I describe the essence of the difficulty, highlighting both the technical issues and the importance of the “plug-and-play property”. I then illustrate some effective approaches to efficient inference based on such models. I conclude by sketching promising new developments and describing some open problems.

---

Zoom link: 709 120 4849 (pw: 1234)

The objective of the study is to evaluate neural circuitry supporting a cognitive control task, and associated practice-related changes via acquisition of blood oxygenation level dependent (BOLD) signal collected using functional magnetic resonance imaging (fMRI). FMR images are acquired from participants engaged in antisaccade (generating a glance away from a cue) performance at two scanning sessions: 1) pre-practice before any exposure to the task, and 2) post-practice, after one week of daily practice on antisaccades, prosaccades (glancing towards a target) or fixation (maintaining gaze on a target). The three practice groups are compared across the two sessions, and analyses are conducted via the application of a model-free clustering technique based on wavelet analysis. This series of procedures is developed to address analysis problems inherent in fMRI data and is composed of several steps: data aggregation, no trend test, decorrelation, principal component analysis and K-means clustering. Also, we develop a semiparametric approach under shape invariance to quantify and test the differences in sessions and groups using the property that brain signals from a task-related experiment may exhibit a similar pattern in regions of interest across participants. We estimate the common function with local polynomial regression and estimate the shape invariance model parameters using evolutionary optimization methods. Using the proposed approach, we compare BOLD signals in multiple regions of interest for the three practice groups at the two sessions and quantify the effects of task practice in these groups.

I will talk about data science and Big Data, and how I view statistics in the data science and Big Data era. Next, I will briefly introduce my research areas in statistics. Finally, I will present some of my interdisciplinary research on functional magnetic resonance imaging data analysis.

---

Direct ZOOM link

While the presence of immune cells within solid tumours was initially viewed positively, as the host fighting to rid itself of a foreign body, we now know that the tumour can manipulate immune cells so that they promote, rather than inhibit, tumour growth. Immunotherapy aims to correct for this by boosting and/or restoring the normal function of the immune system. Immunotherapy has delivered some extremely promising results. However, the complexity of the tumour-immune interactions means that it can be difficult to understand why one patient responds well to immunotherapy while another does not. In this talk, we will show how mathematical, statistical and topological methods can contribute to resolving this issue and present recent results which illustrate the complementary insight that different approaches can deliver.

---

Zoom link: 709 120 4849 (pw: 1234)

A particularly important substructure in modeling joint linear chance-constrained programs with random right-hand sides and finite sample space is the intersection of mixing sets with common binary variables (and possibly a knapsack constraint). In this talk, we first explain basic mixing sets by establishing a strong and previously unrecognized connection to submodularity. In particular, we show that mixing inequalities with binary variables are nothing but the polymatroid inequalities associated with a specific submodular function. This submodularity viewpoint enables us to unify and extend existing results on valid inequalities and convex hulls of the intersection of multiple mixing sets with common binary variables. Then, we study such intersections under an additional linking constraint lower bounding a linear function of the continuous variables. This is motivated from the desire to exploit the information encoded in the knapsack constraint arising in joint linear CCPs via the quantile cuts. We propose a new class of valid inequalities and characterize when this new class along with the mixing inequalities are sufficient to describe the convex hull. This is based on joint work with Fatma Fatma Kılınç-Karzan and Simge Küçükyavuz.

---

https://youtube.com/c/ibsdimag

Our current approach to cancer treatment has been largely driven by finding molecular targets, those patients fortunate enough to have a targetable mutation will receive a fixed treatment schedule designed to deliver the maximum tolerated dose (MTD). These therapies generally achieve impressive short-term responses, that unfortunately give way to treatment resistance and tumor relapse. The importance of evolution during both tumor progression, metastasis and treatment response is becoming more widely accepted. However, MTD treatment strategies continue to dominate the precision oncology landscape and ignore the fact that treatments drive the evolution of resistance. Here we present an integrated theoretical/experimental/clinical approach to develop treatment strategies that specifically embrace cancer evolution. We will consider the importance of using treatment response as a critical driver of subsequent treatment decisions, rather than fixed strategies that ignore it. We will also consider using mathematical models to drive treatment decisions based on limited clinical data. Through the integrated application of mathematical and experimental models as well as clinical data we will illustrate that, evolutionary therapy can drive either tumor control or extinction using a combination of drug treatments and drug holidays. Our results strongly indicate that the future of precision medicine shouldn’t be in the development of new drugs but rather in the smarter evolutionary, and model informed, application of preexisting ones.