Thông báo đặc biệt

Tin nghiên cứu, seminar
Tutorial Seminar on Unsupervised HoloGAN for Neural Rendering
14/06/2019

The computer graphics rendering pipeline is designed for generating high-quality 2D images from 3D scenes, with most research focusing on simulating elements of the physical world, such as light transport models or material simulation. The pipeline, however, can be time-consuming (for example, using ray tracing), and more importantly, it is not differentiable, making it hard to apply for inverse rendering tasks. Computer vision investigates the inference of scene properties from 2D images, and has recently been achieving great success with the adoption of convolution neural networks (CNNs). However, these methods make few explicit assumptions about the physical world or how images are formed from it, and therefore still struggle in tasks that require 3D understanding such as novel-view synthesis, re-texturing or relighting.

 

About speaker: Thu Nguyen-Phuoc is an architect student turned to a roboticist and ML researcher. She's now working on exciting topics of using Deep Unsupervised Learning (HoloGAN) to perform Computer Graphics Rendering (RenderNet, NeurIPS 2019). Join us to learn from her in this Saturday's tutorial seminar. In this tutorial talk, the speaker will present her group's recent work on combing the expressiveness of CNNs and the knowledge of the physical world in the tasks of rendering and inverse rendering.

(To well-prepare for the tutorial seminar attendants are encouraged to read about CNNs, GANs, Rendering in Computer Graphics, and 2 papers by the speaker: RenderNet & HoloGAN at https://www.monkeyoverflow.com)

Địa điểm: phòng I23.

Thời gian: sáng thứ 7, ngày 15/06/2019, từ 9h-12h30.

 
Đọc tiếp...
 
Lớp Hè 2019 (cập nhật)
08/06/2019

Lớp Hè 2019

Các lớp học ngắn với giảng viên thỉnh giảng là hoạt động hè truyền thống ở Khoa Toán - Tin học trường Đại học Khoa học Tự nhiên TPHCM.

Đối tượng tham dự

Các lớp học không thu phí và không tính tín chỉ, mở cho mọi người quan tâm. Người hoàn thành lớp học được nhận một giấy chứng nhận nếu có nhu cầu.

Người muốn dự cần đăng kí bằng cách điền vào mẫu trực tuyến dưới đây:
Mẫu đăng kí học các lớp hè
Danh sách người đã đăng kí

Danh sách lớp

1 An Introduction to Machine Learning: Methodologies and Practical Implementation

  • Đối tượng người học - Audience: anybody in sciences and engineering, including anybody interested in Machine Learning, applied math, life sciences, computer science, electrical engineering, bioengineering, etc, including people outside of academia.
  • Kiến thức cần có - Background: calculus, linear algebra, proficiency with some programming (python, MATLAB)
  • Giảng viên:
    • Kevin Flores, Assistant Professor, Department of Mathematics, Center for Research in Scientific Computation, North Carolina State University
    • Erica Rutter, Postdoctoral fellow, Department of Mathematics, Center for Research in Scientific Computation, North Carolina State University
    • Hien Tran, Alumni Distinguished Graduate Professor, Director Center for Research in Scientific Computation, Department of Mathematics, North Carolina State University
  • Thời gian: June 17--21, 2019. Phòng học: Giảng đường 1, từ chiều Thứ ba: B11A
  • Tóm tắt:
    • What is Machine Learning? “Machine learning teaches computers to do what come naturally to humans and animals: learn from experience. Machine learning algorithms use computational methods to “learn” information directly from data without relying on a predetermined equation as a model. The algorithms adaptively improve their performance as the number of samples available for learning increases” (from MathWorks, Natick, MA)
    • In this mini-lecture series, we will explore the core machine learning concepts and their computational implementation. Several real data, where applicable, will be used to test the numerical implementation.
    • Prerequisite: Participants need to bring their own laptops. Programming proficiency (Python, MATLAB)
  • Course Timeline: Morning Lecture: 9:00 – 10:15 Break: 10:15 – 10:45 Lecture: 10:45 – 12:00 Afternoon Lecture: 2:00 – 3:15 Break: 3:15 – 3:45 Lecture: 3:45 – 5:00
  • Course Topics:
    •  Monday:  ◦ AM (Hien): Bayesian Classifiers, Perceptron, Multilayer Perceptron (Neural Networks)
              ◦ PM (Erica): TensorFlow, UCI Machine Learning Repository: Iris and Diabetes Data Sets
    • Tuesday:  ◦ AM (Hien): Support Vector Machines (separable classes, nonseparable classes), Decision Tree
              ◦ PM (Erica): Tutorial: Scikit-Learn, Credit Card Approval Data Sets
    • Wednesday: (free day) No Lecture
    • Thursday: ◦ AM (Kevin): Convolutional Neural Networks, Computer Vision, Segmentation and Classification
              ◦ PM (Erica): MNIST Data Set, Imagenet, ISBI cell segmentation
    • Friday: ◦ AM (Kevin): Recurrent Neural Networks (RNN), Long Short-Term Memory (LSTM) Networks, Natural Language Processing (NLP), Time Series Data
              ◦ PM (Erica): Time series data, sentiment classification, NLP for translation
              ◦ Optional: Model Evaluation (confusion matrix, loss function, ROC, hypothesis testing), Multi-Class, Dimensionality Reduction (TSNE) – MNIST data last layer dimension reduction

2 Stochastic Models in Ecologie and Evolution: Pure jump Markov Processes in Continuous Time

  • Giảng viên: Sylvie Méléard, Professor, CMAP, École Polytechnique, France
  • Thời gian: July 15–19, 2019. Monday 9:00 - 11:45, Tuesday - Friday 8:45 - 11:30. Problem session by Dr. Hoàng Văn Hà: Monday, Wednesday, Friday, 13:30 - 16:00.
  • Tóm tắt: In these lectures, we will give the structure of the pure jump Markov processes with countable values. The prototype is the Poisson process that we will study in details. We will define the infinitesimal generator and prove the Kolmogorov equations.  Then we will study two well-known examples, useful especially for applications in biology:  branching processes and birth and death processes describing population dynamics. In both cases, we will give criteria of existence and extinction. Finally, we will study  approximations of large populations, showing how these jump processes can be approximated in this case, either by dynamical systems or by stochastic differential equations.
  • Tài liệu tham khảo:

    [1] Modèles aléatoires en écologie et évolution, Mathématiques et Applications 77, SMAI. Springer, 2016. (In French)

    Link download: http://www.cmap.polytechnique.fr/IMG/pdf/LIVRE07102013.pdf

    [2] L.J.S. Allen. An Introduction to Stochastic Processes with Applications to Biology, Second edition. CRC Press, Chapman & Hall/CRC, 2011.

    [3]  V. Bansaye, S. Méléard. Stochastic Models for Structured Populations. Mathematical Biosciences Institute Lecture Series 1.4. Springer 2015.

    Link download: https://arxiv.org/abs/1506.04165

    [4] Ross, Sheldon M. "Stochastic Processes. John Wiley& Sons." New York (1996).

3 An Introduction to Geometric Group Theory

  • Giảng viên: Nhan-Phu Chung, Department of Mathematics, Sungkyunkwan University, Korea.
  • Thời gian: August 12-14, 2019, 9:00-11:00.
  • Tóm tắt: I will present finitely generated groups as viewpoints of metric spaces via word lengths and quasi-isometries. In this part, I will prove a result of Schwarzc-Milnor which is a fundamental observation of geometric group theory. In the second part of the course, I will introduce growth types of finitely generated groups. A landmark result in geometric group theory is Gromov’s theorem stating that a finitely generated group is virtually nilpotent if and only if it has polynomial growth. If time allows I will provide a sketch of Gromov’s proof.
  • References
    1. Tullio Ceccherini-Silberstein and Michel Coornaert, Cellular automata and groups, Springer Monographs in Mathematics, Springer-Verlag, Berlin, 2010.
    2. Pierre de la Harpe, Topics in geometric group theory, Chicago Lectures in Mathematics, University of Chicago Press, Chicago, IL, 2000.
    3. Mikhael Gromov, Groups of polynomial growth and expanding maps, Inst. Hautes Études Sci. Publ. Math. 53 (1981), 53–73.
Đọc tiếp...
 
Seminar Xác suất Thống kê
05/06/2019
Seminar Xác suất Thống kê

Tiến sĩ Nguyễn Hải Đăng, Đại học Alabama sẽ có một buổi báo cáo ngắn về đề tài xác suất (xem chi tiết bên dưới)
và mong muốn gặp gỡ các sinh viên, giảng viên để trao đổi về học bổng tiến sĩ tại Alabama.

  • Tiêu đề: Some New Techniques for Recurrence and Stability of Diffusion Processes in a Random Environment
  • Thời gian:  14 giờ chiều thứ sáu 7/6/2019.
  • Điạ điểm: Bộ môn xác suất Thống kê, Phòng F12, 227 Nguyễn Văn Cừ, ĐHKHTN Tp HCM
  • Tóm tắt: This work focuses on recurrence, ergodicity, and stability of (functional) switching diffusion consisting of continuous and discrete components, in which the discrete component takes values in a countably infinite set. Delays are also allowed. We introduce new techniques by using the log-Laplace transformation, Dupire’s functional Ito formula and coupling methods. Then sufficient conditions for recurrence, ergodicity and stability of the solution process are given. Another distinctive feature is that the path-wise rate of convergence is estimated when the solution is asymptotically stable. Stabilization problems using stochastic time-delayed feedback control can also be examined.  
  • Ngôn ngữ: Tiếng Việt.
Đọc tiếp...
 
Thông báo Seminar Bộ môn Tối Ưu ngày 02.06.2019
30/05/2019
 Kính mời quý thầy cô, nghiên cứu sinh, học viên cao học các sinh viên quan tâm sắp xếp thời gian đến tham dự buổi Seminar Lý thuyết Tối ưu với nội dung sau:

Báo cáo : General versions of the Ekeland variational principle and the Simon satisficing principle.

 Người trình bày: NCS. Lê Phước Hải (Đại học Khoa học Tự nhiên TPHCM)
    Thời gian: Chủ nhật - 02.06.2019 - 09h30
    Địa điểm: phòng F.304 - ĐH Khoa học Tự nhiên, 227 Nguyễn Văn Cừ, Quận 5, TpHCM

Abstract We prove  new general versions of the Ekeland variational principle in a partial quasi-metric space. Unlike the existing versions of this principle, besides a perturbation in terms of the partial quasi-metric, another perturbation being a distance-like bifunction is involved. The classical assumptions on lower semicontinuity of the function and completeness of the space are slightly weakened. The proof technique is new, based onshowing the existence of so-called Ekeland points, which are defined through these two perturbations. Then, we show how these new versions provide striking models for satisficing processes where agents, at each period, do not optimize, but, instead, searchand satisfice. Our new versions of the Ekeland variational principle then are used to develop the Simon satisficing principle which advocates that agents set a satisficing threshold level and search for an alternative until it exceeds this given threshold level. Using a recent variational rationality approach of human dynamics, these new versionsof the Ekeland vatiational principle show the existence of variational traps, such that, starting from an initial position, an agent can satisfice in a worthwhile way, and, being there, is unable to satisfice again in a worthwhile way, giving the end of satisficing process. Moreover, the variatonal approach shows that the Ekeland points represent variational traps (at the intersection of two remarkable sets), which appear to be a set of worthwhile moves starting from the initial point and a set of potential ends (stationary traps).
 
Gặp gỡ Mùa hè 2019, 27--28/7/2019
28/05/2019
"Gặp gỡ Mùa hè" là một cuộc gặp toán học hằng năm kể từ năm 2008 do cựu sinh viên Khoa Toán-Tin học Đại học Khoa học Tự nhiên Thành phố Hồ Chí Minh đang làm toán ở nước ngoài chủ trì.
"Summer Meeting" is an annual mathematical meeting since 2008 organized primarily by alumni of the Faculty of Mathematics and Computer Science, Vietnam National University Ho Chi Minh City-University of Science, who are doing mathematics abroad.

Summer Meeting 2019

The 2019 Meeting is scheduled to hold on Saturday 27 July and Sunday 28 July 2019.

List of speakers

  • Chang Heon Kim, Sungkyunkwan University, Korea
  • Soonhak Kwon, Sungkyunkwan University, Korea
  • Linh Viet Nguyen (Nguyễn Việt Linh), University of Idaho, USA
  • Loc Hoang Nguyen (Nguyễn Hoàng Lộc), University of North Carolina Charlotte, USA
  • Phuc Cong Nguyen (Nguyễn Công Phúc), University of Lousiana Baton Rouge, USA
  • Van Tien Nguyen (Nguyễn Văn Tiên), New York University Abu Dhabi, UAE
  • Trung Tan Nguyen (Nguyễn Tấn Trung), VNUHCM-University of Science, Ho Chi Minh City, Vietnam
  • Hoang Anh Tran (Trần Anh Hoàng), Oak Ridge National Laboratory, USA
  • Son Nguyen Thai Tu (Từ Nguyễn Thái Sơn), University of Wisconsin Madison, USA
  • Son Phung Truong Van (Văn Phụng Trường Sơn), Carnegie Mellon University, USA

Program committee

  • Chung Nhân Phú (Sungkyunkwan University, Korea) ( Địa chỉ email này đang được bảo vệ khỏi chương trình thư rác, bạn cần bật Javascript để xem nó )
  • Huỳnh Quang Vũ (VNUHCM-US) ( Địa chỉ email này đang được bảo vệ khỏi chương trình thư rác, bạn cần bật Javascript để xem nó )
  • Nguyễn Tiến Khải (North Carolina State University, USA) ( Địa chỉ email này đang được bảo vệ khỏi chương trình thư rác, bạn cần bật Javascript để xem nó )
  • Trần Vĩnh Hưng (University of Wisconsin – Madison, USA) ( Địa chỉ email này đang được bảo vệ khỏi chương trình thư rác, bạn cần bật Javascript để xem nó )

Local organizing committee

  • Nguyễn Lê Hoàng Anh (VNUHCM-US)
  • Võ Đức Cẩm Hải (VNUHCM-US)

Supported by

Anybody interested is welcomed to attend. There is no registration fee. For registration, program, and further information: http://www.math.hcmus.edu.vn/summer_meeting

Poster

 

poster-Summer Meeting-2019-ver5

Đọc tiếp...
 
<< Bắt đầu < Trước 1 2 3 4 5 6 7 8 9 10 Tiếp > Cuối >>

Khoa Toán - Tin học, Trường Đại học Khoa học Tự nhiên, Đại học Quốc gia TP Hồ Chí Minh.
Phòng F.009, cơ sở 227 Nguyễn Văn Cừ, Quận 5, TP HCM.