A5103147084- Parallel Computing and Optimization Techniques
- Matrix Theory and Algorithms
- Stochastic Gradient Optimization Techniques
- Advanced Data Storage Technologies
- Privacy-Preserving Technologies in Data
- Electromagnetic Scattering and Analysis
- VLSI and Analog Circuit Testing
- Topic Modeling
- Recommender Systems and Techniques
- Brain Tumor Detection and Classification
- Iterative Methods for Nonlinear Equations
- Environmental Impact and Sustainability
- Cloud Computing and Resource Management
- Data Stream Mining Techniques
- Water Resources and Management
- Advanced Neural Network Applications
- Air Quality Monitoring and Forecasting
- Climate change and permafrost
- Sparse and Compressive Sensing Techniques
- Advanced Memory and Neural Computing
- Environmental and Agricultural Sciences
- Machine Learning and Data Classification
- Anomaly Detection Techniques and Applications
- Atmospheric and Environmental Gas Dynamics
- Data Quality and Management
State Grid Corporation of China (China)
2024-2025
Hohai University
2023
SurveyMonkey (United States)
2023
University of Warwick
2019
The Ohio State University
2018-2019
Nanjing University of Posts and Telecommunications
2006
This study focused on the dynamic relationships among industrial development, energy consumption, economic growth, and carbon emissions in China, with goal of achieving long-term ecological sustainability. Using Panel Vector Autoregressive (PVAR) model Generalized Method Moments (GMM) estimation, panel data from 30 Chinese provinces between 2017 2021 were analyzed. The impulse response analysis variance decomposition demonstrated that subsystems significantly influenced emissions, while...
Generative Language Models have shown promising results in various domains, and some of the most successful applications are related to "concept expansion", which is task generating extensive text based on concise instructions provided through a "seed" prompt. In this presentation we will discuss recent work conducted by Data Science team at SurveyMonkey, where recently introduced new feature that harnesses AI models streamline survey design process. With users can effortlessly initiate...
Many sequential loops are actually recurrences and can be parallelized across iterations as scans or reductions. efforts over the past 2+ decades have focused on parallelizing such by extracting exploiting hidden scan/reduction patterns. These approaches largely been based a heuristic search for closed-form composition of computations loop iterations.
Abstract In this paper, we investigate the influential factors that impact on performance when tasks are co-running a multicore computers. Further, propose machine learning-based prediction framework to predict of tasks. particular, two frameworks developed for types task in our model: repetitive (i.e., arrive at system repetitively) and new submitted first time). The difference between which is have historical running information while do not prior knowledge about Given limited tasks, an...
Communication overhead is a well-known performance bottleneck in distributed Stochastic Gradient Descent (SGD), which popular algorithm to perform optimization large-scale machine learning tasks. In this work, we propose practical and effective technique, named Adaptive Periodic Parameter Averaging, reduce the communication of SGD, without impairing its convergence property.
This study innovatively develops a multi-objective Markal-Macro model, which simultaneously considers three objectives: minimizing carbon emissions from energy consumption, production processes, and maximizing societal welfare. Based on the Cobb–Douglas function, we construct function of emission use it as coupling equation model (Markal is abbreviation market allocation, Macro macroeconomy). enables endogenous variables those related to By collecting relevant data outputs, key economic...
Many sequential loops are actually scans or reductions and can be parallelized across iterations despite the loop-carried dependences. In this work, we consider parallelization of such scan/reduction loops, propose a practical runtime approach called sampling-and-reconstruction to extract hidden patterns in these loops.