Filters








539 Hits in 3.8 sec

Complaint-Driven Training Data Debugging at Interactive Speeds

Lampros Flokas, Weiyuan Wu, Yejia Liu, Jiannan Wang, Nakul Verma, Eugene Wu
2022 Proceedings of the 2022 International Conference on Management of Data  
Recently, Rain [44] proposed complaintdriven data debugging which uses user-specified errors in the output of inference queries (Complaints) to rank erroneous training examples that most likely caused  ...  Combining these optimizations in Rain++ ensures interactive debugging latencies (∼1ms) on models with millions of parameters.  ...  Complaint-Driven Training Data Debugging at Interactive Speeds. In Proceedings of the 2022 International Conference on Management of Data (SIGMOD '22), June 12-17, 2022, Philadelphia, PA, USA.  ... 
doi:10.1145/3514221.3517849 fatcat:3zlb6so7a5cuximdsczonf2aai

Integrating Preventive Maintenance within the Product Life Cycle

Bindu Goel
2011 International Journal of Computer Applications  
It is suggested at this point that the preventive maintenance is done on scheduled basis while perfective maintenance can be user driven.  ...  The reasons to this can be as The Corrective maintenance is error-driven, i.e., this activity is initiated by bugs and equivalent to debugging but it starts only when the system is operational.  ... 
doi:10.5120/3053-4156 fatcat:rzk4umtocbagfcvlsjtln6dkry

From Cleaning before ML to Cleaning for ML

Felix Neutatz, Binger Chen, Ziawasch Abedjan, Eugene Wu
2021 IEEE Data Engineering Bulletin  
What is needed is an end-to-end application-driven approach towards Cleaning For ML, that can leverage signals throughout the entire ML application to optimize the cleaning for application goals and to  ...  Data cleaning is widely regarded as a critical piece of machine learning (ML) applications, as data errors can corrupt models in ways that cause the application to operate incorrectly, unfairly, or dangerously  ...  Debugging systems, such as Rain, consider the output of an entire workflow to trace-back training instances that lead to wrong results.  ... 
dblp:journals/debu/NeutatzCA021 fatcat:thspsxnq4rdx3psrijelwbg6pq

Interactive Semi-automated Specification Mining for Debugging: An Experience Report

Mohammad Jafar Mashhadi, Taha R. Siddiqui, Hadi Hemmati, Howard Loewen
2019 Information and Software Technology  
Conclusion: The main conclusion of this study is that the idea of an "interactive" specification mining rather than a fully automated mining tool is NOT impractical and indeed is useful for the debugging  ...  The results show that a) our approach is feasible, b) it can be automated to some extent, and c) brings advantages over only using their code-level debugging tools.  ...  For instance, if X and Y are speed and takeOffSpeed, respectively, there will be three constraints generated as speed < takeOffSpeed, speed == takeOffSpeed, and speed > takeOffSpeed.  ... 
doi:10.1016/j.infsof.2019.05.002 fatcat:x6qoh5l5c5hxbfegzubxgrcrlm

Conquering the complexity mountain: Full-stack computer architecture teaching with FPGAs

A. Theodore Markettos, Simon W. Moore, Brian D. Jones, Roy Spliet, Vlad A. Gavrila
2016 2016 11th European Workshop on Microelectronics Education (EWME)  
We recount our experiences with rapidly bringing students up to speed with the modern world of computing systems, and some of the lessons we, as course designers, were taught by the process. 978-1-4673  ...  That is in marked contrast to previous years where FPGA build times were the primary complaint and students could waste weeks debugging.  ...  To encourage test-driven development and to speed up application writing, we built an emulator of the peripherals on our FPGA board.  ... 
doi:10.1109/ewme.2016.7496457 dblp:conf/ewme/MarkettosMJSG16 fatcat:wjzgwlaf2vem5orhdhgygsza4u

Struggles of new college graduates in their first software development job

Andrew Begel, Beth Simon
2008 ACM SIGCSE Bulletin  
We report on a "fly-on-thewall" observational study of eight recent college graduates in their first six months of a software development position at Microsoft Corporation.  ...  This is one of the most ubiquitous misconceptionslikely driven by the lack of team-based development and the deadline-driven grading system of academia.  ...  Subject W experienced this in a debugging context.  ... 
doi:10.1145/1352322.1352218 fatcat:jlv5pc7pcfasfkgk5hkcc4vdhe

Struggles of new college graduates in their first software development job

Andrew Begel, Beth Simon
2008 Proceedings of the 39th SIGCSE technical symposium on Computer science education - SIGCSE '08  
We report on a "fly-on-thewall" observational study of eight recent college graduates in their first six months of a software development position at Microsoft Corporation.  ...  This is one of the most ubiquitous misconceptionslikely driven by the lack of team-based development and the deadline-driven grading system of academia.  ...  Subject W experienced this in a debugging context.  ... 
doi:10.1145/1352135.1352218 dblp:conf/sigcse/BegelS08 fatcat:emhs3x3w5fekba6lgej3k3xnde

Virtual instruments: object-oriented program synthesis

K. S. Bhaskar, J. K. Pecol, J. L. Beug
1986 Conference proceedings on Object-oriented programming systems, languages and applications - OOPLSA '86  
Debugging is mndeless and can happen at any time: there is no need to "compile with the debug flag" or to perform any other ritual.  ...  ,ieve at least an order of magnitude speed-up to be essential for it to be usable. We decided therefore, not to attempt to drive hardware with this implementation.  ... 
doi:10.1145/28697.28727 dblp:conf/oopsla/BhaskarPB86 fatcat:pon7cmnwmvhfxmtwzijrsed7uy

A Worldwide Look Into Mobile Access Networks Through the Eyes of AmiGos [article]

Matteo Varvello, Yasir Zaki
2022 arXiv   pre-print
The goal is to answer high level questions like "which performance can you expect at a 10cost around the world?", as well as low level questions like "what is the culprit of a slowdown?  ...  Specifically: a) 50 of encountering a low data rate, b) only 20 characterized by low latencies (<20ms to popular domains), c) networks across Asia, central/south America, and Africa have much higher CDN  ...  Overall, our daily limit was very conservative since the median total data consumption was close to 6 GB; this is driven by how mobile users are since at home they connected the AmiGo measurement endpoints  ... 
arXiv:2209.04129v1 fatcat:zdavavdvkfdifb3hep6u56kkpu

Operationalizing Machine Learning: An Interview Study [article]

Shreya Shankar, Rolando Garcia, Joseph M. Hellerstein, Aditya G. Parameswaran
2022 arXiv   pre-print
The process of operationalizing ML, or MLOps, consists of a continual loop of (i) data collection and labeling, (ii) experimentation to improve ML performance, (iii) evaluation throughout a multi-staged  ...  Work was done while Hellerstein was on leave at Sutter Hill Ventures.  ...  build the infrastructure 16 8 5 data pipeline data iteration, fresh data 15 8 6 fast & simple high iteration speed, agile, rapid cycles 24 7 7 production bugs debugging and bugs 18 7 8 tests AB Testing  ... 
arXiv:2209.09125v1 fatcat:dlqmvgehbfc67k5q2al3gtwiqa

Interactions with big data analytics

Danyel Fisher, Rob DeLine, Mary Czerwinski, Steven Drucker
2012 Interactions  
The Nature of Analytics Work The term analytics (including its big data form) is often used broadly to cover any data-driven decision making.  ...  Doubling the memory or computation speed of a machine, however, does not double its speed; it can impose non-linear costs as communication overhead, storage, and other aspects change.  ...  About thE Authors Danyel Fisher is a researcher at Microsoft Research working on information visualization. In his research he looks for new ways to explore and interact with data.  ... 
doi:10.1145/2168931.2168943 fatcat:6pb3rcozyjewlj2gtwpinmsxzm

QFix

Xiaolan Wang, Alexandra Meliou, Eugene Wu
2017 Proceedings of the 2017 ACM International Conference on Management of Data - SIGMOD '17  
Data-driven applications rely on the correctness of their data to function properly and effectively.  ...  While data cleaning tools can purge datasets of many errors before the data is used, applications and users interacting with the data can introduce new errors.  ...  speeds.  ... 
doi:10.1145/3035918.3035925 dblp:conf/sigmod/WangM017 fatcat:vmcsvtm2dfg4vhauzoczv76ywe

An Approach for iOS Applications' Testing

Ivans Kuļešovs, Juris Borzovs, Aigars Susters, Vineta Arnicane, Guntis Arnicans, Kirils Keiduns, Janis Skutelis
2018 Baltic Journal of Modern Computing  
Solutions aimed at overcoming the limitations of out of the box UIAutomation are united in tTap framework developed by the authors.  ...  Tests can be automated at various levels.  ...  They are: • Analytic School; • Standard School; • Quality School; • Context-Driven School; • Agile School.  ... 
doi:10.22364/bjmc.2018.6.1.05 fatcat:eu6cxwpq3vctjjhp72mgor6ly4

Using Students as Experimental Subjects in Software Engineering Research -- A Review and Discussion of the Evidence [article]

Dror G. Feitelson
2015 arXiv   pre-print
The earliest is the study by Sackman et al. from 1968 regarding interactive debugging, which showed that both trainees and experienced professionals benefited from using an interactive setting [151] .  ...  At the very least, appropriate training should be provided, and the level of proficiency should be assessed.  ... 
arXiv:1512.08409v1 fatcat:yjhik5uk6rh7plsk6evwnxpciy

MapReduce is Good Enough? If All You Have is a Hammer, Throw Away Everything That's Not a Nail! [article]

Jimmy Lin
2012 arXiv   pre-print
From this combined perspective I reflect on the current state and future of "big data" research.  ...  Hadoop is currently the large-scale data analysis "hammer" of choice, but there exist classes of algorithms that aren't "nails", in the sense that they are not particularly amenable to the MapReduce programming  ...  To the extent that we can provide tools to allow rich, interactive, incremental interactions with large data sets, we can boost the productivity of data scientists, thereby increasing their ability to  ... 
arXiv:1209.2191v1 fatcat:yu4z5pwryjghtaspm6qjvyna4y
« Previous Showing results 1 — 15 out of 539 results