Upgrade to Pro — share decks privately, control downloads, hide ads and more …

End-to-end Quality Control in Approximate Computing

3c332dfc0b438785cb10c5234652dd66?s=47 Yuhao Zhu
March 24, 2016
Education
0
98

End-to-end Quality Control in Approximate Computing

WAX 2016 co-located with ASPLOS 2016

3c332dfc0b438785cb10c5234652dd66?s=128

Yuhao Zhu

March 24, 2016
Tweet

More Decks by Yuhao Zhu

See All by Yuhao Zhu
RTNN: Accelerating Neighbor Search Using Hardware Ray Tracing
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
93
Architecture Support for Robust Deep Learning: Exploiting Software 1.0 Techniques to Defend Software 2.0
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
120
Watt-Wise Web: Architecting a Responsive and Energy-Efficient Mobile Web
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
21
Tail Latency in Node.js: Energy Efficient Turbo Boosting for Long Latency Requests in Event-Driven Web Services
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
190
Algorithm-SoC Co-Design for Mobile Continuous Vision
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
270
Energy and Power Measurement on Mobile Devices
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
85
Defense talk: Energy-Efficient Mobile Web Computing
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
380
GreenWeb: Language Extensions for Energy-Efficient Mobile Web Computing
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
180
The Human Processing Unit (HPU) as a New Approximate Computing Substrate
3c332dfc0b438785cb10c5234652dd66?s=48 queeten
0
170

Other Decks in Education

See All in Education
View Manipulation and Reduction - Lecture 9 - Information Visualisation (4019538FNR)
1135dc242dcff3b90ae46fc586ff4da8?s=48 signer
PRO
0
510
APERITIVO ITALIANO A1.0
C548b8f4f64fdf90c7d62b22d23e127e?s=48 institucional
0
220
読書に悩むあなたに贈る50の読書方法カタログ / reading catalog Best50
055e63a40190884fdc17b918c6f6aa7e?s=48 aki_moon
9
2.2k
WindowsコンテナDojo 全体スケジュールのご案内
B876625b5b40b0621fec49b07bd90e25?s=48 utsukibm
0
410
ITエンジニアも経営スキルを身につけよう 第1回
227e9b3b0bf93a6c52874004a3ae20eb?s=48 lsuzuki
1
240
Tangible, Embedded and Embodied Interaction - Lecture 9 - Next Generation User Interfaces (4018166FNR)
1135dc242dcff3b90ae46fc586ff4da8?s=48 signer
PRO
0
660
Data Management and Analytics Specialisation
1135dc242dcff3b90ae46fc586ff4da8?s=48 signer
PRO
0
430
H5P-työkalut
18cdb6f294c8cb10a46167e462c2a51f?s=48 matleenalaakso
2
19k
ITエンジニアも経営スキルを身につけよう 第3回
227e9b3b0bf93a6c52874004a3ae20eb?s=48 lsuzuki
0
180
My toolbox is full of shiny tools, do I also need super powers?
81689b093f75cf3f383e581ca57188df?s=48 minecr
0
130
Moodle 4.0 : les nouveautés avec Pimenko
14b832a45ca90251ff3a52ac1daaed99?s=48 pimenko
0
500
World War I and the Red Scare
9d495f5d79fbbc9ddf3100da74f986c9?s=48 oripsolob
0
690

Featured

See All Featured
Large-scale JavaScript Application Architecture
96270e4c3e5e9806cf7245475c00b275?s=48 addyosmani
499
110k
Designing Experiences People Love
1b75d89772b7eea1f6c89fbf362607d9?s=48 moore
130
22k
KATA
E9221b172e78e888355fa2fe4541b595?s=48 mclloyd
7
8.8k
Templates, Plugins, & Blocks: Oh My! Creating the theme that thinks of everything
E195ae45320d9202eaa01c9f1d31a416?s=48 marktimemedia
15
980
Creatively Recalculating Your Daily Design Routine
48c098b6579220a67a6ba949be6e4b5a?s=48 revolveconf
207
10k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
A9a491b0fcbe0fbce3d64063a37add99?s=48 rmw
21
1.4k
Side Projects
027d1ebf66cc039a0bd3b55eeadbe75d?s=48 sachag
450
37k
Why You Should Never Use an ORM
88bc30284c6a424b63a92aad27d0ba36?s=48 jnunemaker
PRO
47
7.6k
Intergalactic Javascript Robots from Outer Space
6804f1775cb4babfcc3851298566fbce?s=48 tanoku
260
25k
Cheating the UX When There Is Nothing More to Optimize - PixelPioneers
F383c6a4dc55e331bbe2987b622cee6b?s=48 stephaniewalter
269
12k
How To Stay Up To Date on Web Technology
8081b26e05bb4354f7d65ffc34cbbd67?s=48 chriscoyier
780
250k
Rails Girls Zürich Keynote
24fc194843a71f10949be18d5a692682?s=48 gr2m
87
12k

Transcript

  1. 1 WAX 2016 Scalable End-to-end Quality Control in Approximate Computing

    Yuhao Zhu UT Austin with Deepti Ghadiyaram, Matt Lease, and Vijay Janapa Reddi

  2. User-centric Application Image compression

  3. User-centric Application Image compression Quality Metric Average Pixel Error Rate

  4. Human Perception User-centric Application Image compression Quality Metric Average Pixel

    Error Rate User Satisfaction

  5. Human Perception User-centric Application Image compression Quality Metric Average Pixel

    Error Rate ? User Satisfaction

  6. Quality Score (QScore) Human Perception User-centric Application Image compression

  7. Quality Score (QScore) Human Perception User-centric Application Quality Assessment Model

    Image compression QScore = f(image) [Wang et al., 2004] [Seshadrinathan et al., 2010]

  8. Quality Score (QScore) Human Perception User-centric Application Quality Assessment Model

    Image compression QScore = f(image) [Wang et al., 2004] [Seshadrinathan et al., 2010]

  9. Human Perception User-centric Application Quality Assessment Model Audio Video Text

    Image QScore Image Model Audio Model Video Model Text Model Audio QScore Video QScore Text QScore Image