蛋播视频一区,无码鲁丝一区二区,精品 久久 五月天,国产老熟女,五月草草在线观看,中文日韩欧美,情色一区二区三区,欧美日韩亚洲激情在线,亚洲制服在线香蕉

運(yùn)籌優(yōu)化與運(yùn)營(yíng)管理研究前沿系列

TITLE:

M-natural/S-Convexity and Applications

ABSTRACT:

Classical monotone comparative statics in parametric optimization problems establishes the existence of nondecreasing optimal solutions (in parameters) through lattice programming. In this talk, we use M-natural-convexity, one of the key concepts in discrete convex analysis, and introduce a new concept of S-convexity as its extension to develop monotone comparative statics in parallel to lattice programming. Specifically, we show that in a parametric minimization model, the optimal solution (set) is nonincreasing (instead of nondecreasing) in the parameters when the objective function is M-natural/S-convex and the constraint is a box, and M-natural/S-convexity is preserved under the optimization operation. We illustrate their applications on several operations models including a multi-product newsvendor model with a joint capacity and a portfolio contract model.

ID: 88555021030

密碼：123456

SPEAKER:

Xin Chen is a professor at the University of Illinois at Urbana-Champaign. He obtained his PhD from MIT in 2003, MS from Chinese Academy of Sciences in 1998 and BS from Xiangtan University in 1995. His research interest lies in optimization, data analytics, revenue management and supply chain management. He received the Informs revenue management and pricing section prize in 2009. He is the coauthor of the book “The Logic of Logistics: Theory, Algorithms, and Applications for Logistics and Supply Chain Management (Second Edition, 2005, & Third Edition, 2014)”, and serving as the department editor of logistics and supply chain management of Naval Research Logistics and an associate editor of several journals including Operations Research, Management Science, Mathematics of Operation Research and Production and Operations Management.

運(yùn)籌優(yōu)化與運(yùn)營(yíng)管理研究前沿系列

TITLE:

Learning Algorithms in OM

ABSTRACT:

This talk is focused on the design and analysis of online learning algorithms for operations management models. Recent research projects will be discussed in greater details, including stochastic inventory systems and revenue management systems.

ID：84547929989

密碼：123456

SPEAKER:

Cong Shi is an associate professor in the Department of Industrial and Operations Engineering at the University of Michigan at Ann Arbor. His main research interests include supply chain management, revenue management, and service operations. He has won the first place in the INFORMS George Nicholson Student Paper Competition, the third place in the INFORMS Junior Faculty Interest Group (JFIG) Paper Competition, and the finalist for the MSOM Data Driven Challenge. He received his Ph.D. in Operations Research from MIT in 2012, and his B.S. in Mathematics from the National University of Singapore in 2007.

信息管理與電子商務(wù)研究前沿系列

TITLE:

Big Data, AI and Financial Economics

ABSTRACT:

I will give three talks on using statistical machine learning techniques and big data to solve problems in finance and economics. I will give an extensive introduction on the interfaces of statistical machine learning, data science and AIs in the first lecture.

Talk 1: Demystifying High Frequency Trading（8月4日）

This paper studies the predictability in ultra high-frequency finance, with focus on the momentum measured by proportions of price changes and trade directions and duration that reflects trading speeds in very short time windows. These predictability issues are investigated by using three measures of time: calendar, trade and volume clocks. Using statistical machine learning methods on complete transaction and quote update data of $101$ stocks in the S\&P 100 index over two full years from 2019 to 2020, we quantified and documented the predictability and confirmed that it exists universally. For a median stock, a $10.5\%$ out-of-sample $R^2$ of 5-second trade returns can be predicted using merely past trade and quote data with about $64\%$ of correctly predicting trade directions. For prediction of 10-trade duration, the median out-of-sample $R^2$ is $9.8\%$. The important predictors are also unveiled. We also investigated how the predictability depends on the market environments. Returns and directions are found more predictable for stocks that have smaller nominal share prices, that are less liquid, less volatile, and less related with the market. In contrast, predictability for durations are higher under liquid and volatile conditions. We also investigated the the timeliness of data and found that predictability resides in the most recent $10$ milliseconds, $10$ transactions or $10$ lots transacted, and decreases sharply once a small delay is introduced. We also simulate the possible ability of high-frequency traders in making short-term and imperfect predictions on future order flow. Such ability, even just correctly predict the sign, is able to boost 5-second return $R^2$ from $14.0\%$ up to $27.1\%$ and direction accuracy from $68.3\%$ up to $79.0\%$. Our study shed light on understanding micro-structure of price evolution and high-frequency finance.

Talk 2: Learning Housing Activeness from Multi-source Big Data（8月5日）

Measuring timely high-resolution socioeconomic outcomes is critical for policy-making and evaluation. Such indices in fine granularity are constructed mostly by extrapolating the existing census data. However, extrapolation creates large approximation errors and census data is costly and noisy due to the subjective response during the interview, which could be more severe in sectors like the real estate and developing countries like China. With the help of machine learning and cheaply available data such as social media and nightlight, it is now possible to predict such indices in fine granularity. This paper demonstrates an adaptive way to measure the time trend and spatial distribution of housing activeness with the help of multiple easily-accessible datasets. We first identified the regional activeness status at the individual level from energy consumption data and then matched it with nightlight and land use data geographically. Then we proposed a Factor-Augmented Regularized Model for Prediction (FarmPredict) that effectively lifts the prediction space and solves the colinearity problem in high-dimensional data. FarmPredict allows us to extend the regional results to the city level, with a $75\%$ out-of-sample explanation of the spatial and timeliness variation in the housing usage. FarmPredict is not only a model but an analytical framework of machine learning on high-dimensional data, showing broad potential applications to other social science problems. Since energy is indispensable for life, our method is highly transferable with the requirement of only public and accessible data. Our paper demonstrates the power of machine learning in understanding socioeconomic outcomes when the census and survey data is costly or unavailable.

Talk3：How Much Can Machines Learn Finance From Chinese Text Data?（8月6日）

Most studies on equity markets using text data focus on English-based specified sentiment dictionaries or topic modeling. However, can we predict the impact of news directly from the text data? How much can we learn from such a direct approach? We present here a new framework for learning text data based on the factor model and sparsity regularization, called FarmPredict, to let machines learn financial returns automatically. Unlike other dictionary-based or topic models that have stringent pre-screening processes, our framework allows the model to extract information more fully from the whole article. We demonstrate our study on the Chinese stock market, as Chinese text has no natural spaces between words and phrases and the Chinese market has a very large proportion of retail investors. These two specific features of our study differ significantly from the previous literature that focuses on English-text and the U.S. market. We validate our method using the literature on the Chinese stock market with several existing approaches. We show that positive sentiments scored by our FarmPredict approach generate on average 83 bps stock daily excess returns, while negative news has an adverse impact of 26 bps on the days of news announcements, where both effects can last for a few days. This asymmetric effect aligns well with the short-sale constraints in the Chinese equity market. As a result, we show that the machine-learned sentiments do provide sizable predictive power with an annualized return of 116% with a simple investment strategy and the portfolios based on our model significantly outperform other models. This lends further support that our FarmPredict can learn the sentiments embedded in financial news. Our study also demonstrates the far-reaching potential of using machines to learn text data.

8月4日

ID：458885434

密碼：123456

8月5日

ID：625242085

密碼：123456

8月6日

ID：475405405

密碼：123456

SPEAKER:

人工智能領(lǐng)域研究前沿系列

TITLE:

Deep Learning and Non-Functional Requirements

ABSTRACT:

Artificial intelligence (AI) is aimed at building machines that can algorithmically learn to group objects, to perform classifications, to recognize patterns, to make inferences, and so forth. In addition to these functional requirements, the AI-based systems shall exhibit certain non-functional properties, such as robustness and explainability. In this talk, I will highlight our ongoing work of engineering and assessing deep learning solutions to conquer one of the most pressing societal problems: combined sewer overflows (CSOs). CSOs have significant global impacts, harming human and environmental health. In the United States, for example, nearly 860 cities and towns have combined sewer systems that manage storm water as well as waste water, causing about 850 billion gallons of untreated water to be discharged into waterways annually. Due to the time-series data collected by our partner organization, we have developed recurrent neural networks (RNNs) to predict the CSO events. Meanwhile, non-functional requirements (NFRs) have played a crucial role in choosing a suitable RNN implementation and eliciting new application scenarios of deep learning. I will share our lessons learned and provoke the discussions about better engineering the deep learning solutions to satisfy the stakeholders’ needs.

ID：88447627559

參會(huì)鏈接: https://us02web.zoom.us/j/88447627559

密碼：626229

SPEAKER:

Nan Niu is an Associate Professor at the Department of Electrical Engineering and Computer Science (EECS), University of Cincinnati, USA. His current research interests focus on requirements engineering in machine learning, continuous deployment, and model-driven contexts. He is also investigating practical and automated ways to manage software traceability and to test scientific software. He is an Associate Editor of Springer’s Requirements Engineering journal, and has published numerous papers in the premier software engineering conferences, such as ICSE, ESEC/FSE, and RE. He is a recipient of the U.S. NSF CAREER Award, two Best Paper Awards, and one Most Influential Paper Award.

上財(cái)信息

講座信息將提前預(yù)告

歡迎校內(nèi)外師生校友旁聽！