the image of me

Vahideh Akhlaghi

PhD Student
Computer Science and Engineering
UC San Diego

About Me

I am currently a 4th year PhD student in the department of Computer Science and Engineering, at UC San Diego. I am working in Microelectronic Embedded System Laboratory, led by Prof. Rajesh Gupta.
I received my Master and Bachelor Degree in Computer Engineering from University of Tehran.

Resume

You can find my resume here.

Current Research

Accelerator Design for Machine Learning Algorithms

Distributed training of Convolutional Neural Networks (CNN) on heterogenous systems

Reducing the computation in deep Convolutional Neural Networks (CNN) training and inference through approximate computing

In-memory implementation of tree-based learning algorithms using resistive memories

Design Automation for Approximate Computing

Machine Learning-based optimization framework for energy-error trade-offs in combining approximate methods in approximate computing

Error-bounded function approximation with probabilistic data structure

Publications

V. Akhlaghi, A. Yazdanbakhsh, K. Samadi, R. K. Gupta, H. Esmaeilzadeh, “SnaPEA: Predictive Early Activation for Reducing Computation in Deep Convolutional Neural Networks”, to appear in International Symposium on Computer Architecture (ISCA), 2018.

V. Akhlaghi, S. Gao, R. K. Gupta, “LEMAX: Learning-based Energy Consumption Minimization in Approximate Computing with Quality Guarantee”, to appear in ACM/IEEE Design Automation Conference (DAC), 2018.

X. Jiao, V. Akhlaghi, Y. Jiang, R. K. Gupta, “Energy-Efficient Neural Networks using Approximate Computation Reuse”, IEEE Design, Automation, and Test in Europe (DATE), 2018.

V. Akhlaghi, A. Rahimi, R. K. Gupta, “Resistive Bloom Filters: From Approximate Membership to Approximate Computing with Bounded Errors”, IEEE Design, Automation and Test in Europe (DATE), 2016.

J. Koh, B. Balaji, V. Akhlaghi, Y. Agarwal, R. K. Gupta, “Quiver: Using Control Perturbations to Increase the Observability of Sensor Data in Smart Buildings. CoRR abs/1601.07260 (2016)

V. Akhlaghi, M. Kamal, A. Afzali-kusha, M. Pedram, “An Efficient Network on-Chip Architecture Based on Isolating Local and non-Local Communications”, Elsevier Journal of Computers and Electrical Engineering, 2015.

V. Akhlaghi, M. Kamal, A. Afzali-kusha, M. Pedram, “An Efficient Network on-Chip Architecture Based on Isolating Local and non-Local Communications”, IEEE Design Automation and Test in Europe (DATE ‘13), 2013.

Poster Presentations

V. Akhlaghi, M. Imani, T. S. Rosing, R. K. Gupta. Acceleration Architecture for Decision Tree Classification Algorithm. ACM/IEEE Design Automation Conference (DAC’17). Work in Progress.

V. Akhlaghi, A. Rahimi, R. K. Gupta, “Resistive Bloom Filters: From Approximate Membership to Approximate Computing with Bounded Errors”, Research Expo at UCSD, 2017.

V. Akhlaghi, H. Sohofi, Z. Navabi, “Migrating from RTL to ESL via SystemC Training”, IEEE/ACM Design Automation and Test in Europe (DATE ‘10), Germany, 2010, University Booth.

Contact Info

Email: vakhlaghi at eng dot ucsd dot edu
Office: 2150 EBU3b
Dept. of Computer Science and Engineering
9450 Gilman Drive, Mail Code 0404
La Jolla, CA 92093-0404