Optimizing big data analytics
Event details
| Date | 21.12.2015 |
| Hour | 10:00 › 12:00 |
| Speaker | Immanuel Trummer, EPFL |
| Location | |
| Category | Conferences - Seminars |
Abstract:
We are collecting data at unprecedented rates. This data contains valuable insights but we need complex analytics to extract them. Our ability to generate data threatens however to outgrow our capability to analyze it. In this context, it becomes crucial to exploit our computational resources in the most efficient way. My research focuses on planning, optimization, and scheduling problems that occur in the context of data analytics. The ultimate goal is to save analytics users from daunting planning and configuration tasks and to exploit computational resources in the optimal way.
In this talk, I will give an overview of my research in that space. I will discuss how the query optimization problem, a core problem in the area of databases, needs to be revisited in the context of big data analytics. The optimization problems that we encounter in big data analytics are often particularly challenging. I will show how techniques such as approximation and massive parallelization can help to tackle them. In the last part of my talk, I will describe my experiences when experimenting on the first commercially available quantum computer, the D-Wave adiabatic quantum annealer. I will show how I used that device to solve a classical analytics related optimization problem and elaborate on the long-term potential of quantum computing for analytics optimization.
Biography:
Immanuel Trummer is a PhD student at EPFL under supervision of Christoph Koch. His research led to various publications at the VLDB and SIGMOD conferences. His papers were invited for publication into the VLDB Journal (“Best of VLDB 2015”) and selected for the ACM SIGMOD Research Highlight Award 2015. He received the European Google PhD Fellowship 2015 in structured data analysis and is alumnus of the German National Academic Foundation.
We are collecting data at unprecedented rates. This data contains valuable insights but we need complex analytics to extract them. Our ability to generate data threatens however to outgrow our capability to analyze it. In this context, it becomes crucial to exploit our computational resources in the most efficient way. My research focuses on planning, optimization, and scheduling problems that occur in the context of data analytics. The ultimate goal is to save analytics users from daunting planning and configuration tasks and to exploit computational resources in the optimal way.
In this talk, I will give an overview of my research in that space. I will discuss how the query optimization problem, a core problem in the area of databases, needs to be revisited in the context of big data analytics. The optimization problems that we encounter in big data analytics are often particularly challenging. I will show how techniques such as approximation and massive parallelization can help to tackle them. In the last part of my talk, I will describe my experiences when experimenting on the first commercially available quantum computer, the D-Wave adiabatic quantum annealer. I will show how I used that device to solve a classical analytics related optimization problem and elaborate on the long-term potential of quantum computing for analytics optimization.
Biography:
Immanuel Trummer is a PhD student at EPFL under supervision of Christoph Koch. His research led to various publications at the VLDB and SIGMOD conferences. His papers were invited for publication into the VLDB Journal (“Best of VLDB 2015”) and selected for the ACM SIGMOD Research Highlight Award 2015. He received the European Google PhD Fellowship 2015 in structured data analysis and is alumnus of the German National Academic Foundation.
Practical information
- General public
- Free
- This event is internal
Contact
- Babak Falsafi