EE Student Information

EE380 Computer Systems Colloquium

EE380 Computer Systems Colloquium: Deep Speech: Scaling up end-to-end speech recognition.

Topic: 
Deep Speech: Scaling up end-to-end speech recognition.
Abstract / Description: 

Speech recognition is still an unsolved problem in AI. Humans transcribe speech substantially better than machines, particularly when the speech is noisy, accented or spoken in a natural, unaffected manner. Over the past half-century slow yet steady progress has been made in speech recognition punctuated with rare breakthroughs including the Hidden Markov Model in the 70s and, more recently, Deep Neural Networks.

In fact, the past few years have witnessed large strides in many machine learning problems including speech recognition and computer vision. This is mostly due to the resurgence of Deep Learning - a class of machine learning algorithms consisting of large neural networks with many layers. Two main drivers of progress in this field have been efficient computation at scale using GPUs and the ability to acquire or construct large labeled datasets. However, as these algorithms continue to scale up, new challenges arise. In particular capturing, annotating and efficiently accessing the data needed to train these algorithms is a resource intensive problem. Furthermore, as the dataset and model sizes continue to increase, efficiently training and evaluating these networks poses a challenge.

In this presentation I will give an overview of the current state of speech recognition technology. I will also discuss the challenges we must overcome in order to make progress and eventually approach human level performance. This presentation will include a high-level introduction to Deep Learning in addition to reviewing some of the latest applications of it. I will focus on Deep Speech, a Deep Learning based speech recognition system built at Baidu Research's Silicon Valley AI lab, which has shown great potential for rapid progress in speech recognition.

Date and Time: 
Wednesday, February 4, 2015 - 4:15pm to 5:15pm
Venue: 
Gates B03

EE380 Computer Systems Colloquium: Computational Epidemiology: The role of big data and pervasive informatics

Topic: 
Computational Epidemiology: The role of big data and pervasive informatics
Abstract / Description: 

Pandemics such as H1N1 influenza are global outbreaks of infectious disease. Human behavior, social contact networks, and pandemics are closely intertwined. The ordinary behavior and daily activities of individuals create varied and dense social interactions that are characteristic of modern urban societies. They provide a perfect fabric for rapid, uncontrolled disease propagation. During the course of an epidemic, individuals and institutions modify their normal behavior based on their perceived severity and risk. The resulting co-evolution of individual and collective behaviors, contact networks and epidemics must be taken into account while designing effective planning and response strategies.

Recent advances in high performance pervasive computing and big data have created new opportunities for collecting, integrating, analyzing and accessing information about evolving social interactions. The advances in network and information science that build on this new capability provide entirely new ways for reasoning and controlling epidemics.

In this talk I will overview of the state of the art in computational networked epidemiology with an emphasis on computational thinking and high performance computing oriented decision-support environments to support planning and response in the event of pandemics. I will describe our approach within the context of a specific recent application: modeling to support Ebola Outbreak Response in West Africa.

Date and Time: 
Wednesday, January 21, 2015 - 4:15pm to 5:30pm
Venue: 
Gates B03

EE380 Computer Systems Colloquium: Nadia Heninger

Topic: 
How not to generate random numbers
Abstract / Description: 

Randomness is essential to cryptography: cryptographic security depends on private keys that are unpredictable to an attacker. But how good are the random number generators that are actually used in practice? In this talk, I will discuss several large-scale surveys of cryptographic deployments, including TLS, SSH, Bitcoin, and secure smart cards, and show that random number generation flaws are surprisingly widespread. We will see how many of the most commonly used public key encryption and signature schemes can fail catastrophically if used with faulty random number generators, and trace many of the the random number generation flaws we encountered to specific implementations and vulnerable implementation patterns.

Date and Time: 
Wednesday, May 13, 2015 - 4:15pm to 5:30pm
Venue: 
Gates B03

EE380 Computer Systems Colloquium: Programming Should Be More Than Coding

Topic: 
Programming Should Be More Than Coding
Abstract / Description: 

Writing a program involves three tasks:

  1. Deciding what the program should do.
  2. Deciding how the program should do it.
  3. Coding: Implementing these decisions in code.

Too often, all three are combined into the process of coding. This talk explains why they should be separated, and discusses how to perform the first two.

Date and Time: 
Wednesday, April 8, 2015 - 4:15pm to 5:30pm
Venue: 
Gates B03

Pages

Subscribe to RSS - EE380 Computer Systems Colloquium