Planning and Executing Practice-Impactful Research
2 likes1,472 views
The document discusses the importance of impactful research within the context of academia-industry collaboration, emphasizing the need for engagement with real-world problems and continuous adoption of academic findings by practitioners. It highlights challenges faced when transitioning research into practice, including differing goals between academia and industry, and proposes strategies for fostering successful partnerships. Additionally, it outlines various levels of practice impact in research and stresses the significance of mutual contributions and long-term relationships in collaborations.
![Desirable Problems for Academia-Industry
Collaborations
• Not all industrial problems are worth effort investment from university
groups
• High business/industry value
• Allow research publications (not business secret) to advance the knowledge
• Challenging problem (does it need highly intellectual university researchers?)
• Desirably real man-power investment from both sides
• My recent examples
• Tencent WeChat [FSE’16 Industry], [ICSE’17 SEIP]: Android app testing/analysis
• Exploring collaborations with Baidu, Alibaba, Huawei
• Exploring new collaborations with MSRA SA](https://image.slidesharecdn.com/practiceimpactresearch2017-taoxie-170806093923/85/Planning-and-Executing-Practice-Impactful-Research-10-320.jpg)
Planning and Executing Practice-Impactful Research
- 1. Planning and Executing Practice-Impactful Research Tao Xie Department of Computer Science University of Illinois at Urbana-Champaign taoxie@illinois.edu http://taoxie.cs.illinois.edu/
- 4. Evolution of Research Assessment • #Papers • #International Venue Papers • #SCI/EI Papers • #CCF A (B/C) Category Papers • ??? CRA 2015 Report: “Hiring Recommendation. Evaluate candidates on the basis of the contributions in their top one or two publications, …” “Tenure and Promotion Recommendation. Evaluate candidates for tenure and promotion on the basis of the contributions in their most important three to five publications (where systems and other artifacts may be included).” http://cra.org/resources/best-practice-memos/incentivizing-quality-and-impact-evaluating-scholarship-in-hiring-tenure-and-promotion/
- 5. Societal Impact ACM Richard Tapia Celebration of Diversity in Computing Join us at the next Tapia Conference in Atlanta, GA on September 20-23, 2017! http://tapiaconference.org/ Margaret Burnett: “Womenomics & Gender-Inclusive Software” “Because anybody who thinks that we’re just here because we’re smart forgets that we’re also privileged, and we have to extend that farther. So we’ve got to educate and help every generation and we all have to keep it up in lots of ways.” – David Notkin, 1955-2013 Andy Ko: “Why the Software Industry Needs Computing Education Research”
- 6. Impact on Research Communities Beyond SE Representational State Transfer (REST) as a key architectural principle of WWW (2000) Related to funding/head-count allocation, student recruitment, … community growth Roy Fielding Richard Taylor … Andreas Zeller Delta debugging (1999)Symbolic execution (1976) also by James King, William Howden, Karl Levitt, et al. Lori Clarke
- 7. Practice Impact • Diverse/balanced research styles shall/can be embraced • Our community already well appreciates impact on other researchers, e.g., SIGSOFT Impact Awards, ICSE MIP, paper citations • But often insufficient effort for last mileage or focus on real problems • Strong need of ecosystem to incentivize practice impact pursued by researchers • Top down: • Bottom up: • Conference PC for reviewing papers • Impact counterpart of “highly novel ideas”? • Impact counterpart of “artifact evaluation”? • Promote and recognize practice impact • Counterpart of ACM Software System Award?
- 8. Practice-Impact Levels of Research • Study/involve industrial data/subjects • Indeed, insights sometimes may benefit practitioners • Hit (with a tool) and run • Authors hit and run (upon industrial data/subjects) • Practitioners hit and run • Continuous adoption by practitioners • Importance of benefited domain/system (which can be just a single one) • Ex. WeChat test generation tool WeChat with > 900 million users • Ex. MSRA SA on SAS MS Online Service with hundreds of million users • Ex. Beihang U. on CarStream Shenzhou Rental with > 30,000 vehicles over 60 cities • Scale of practitioner users • Ex. MSR Pex Visual Studio 2015+ IntelliTest • Ex. MSR Code Hunt with close to 6 million registered/anonymous/API accounts • Ex. MSRA SA XIAO Visual Studio 2012+ Clone Analysis Think about >90% startups fail! It is challenging to start from research and then single-handedly bring it to continuous adoption by target users; Academia-industry collaborations are often desirable.
- 9. Practice-Impact Levels of Research • If there are practice impacts but no underlying research (e.g., published research), then there is no practice-impactful research • More like a startup’s or a big company’s product with business secrets • Some industry-academia collaborations treat university researchers (students) like cheap(er) engineering labor no or little research
- 10. Desirable Problems for Academia-Industry Collaborations • Not all industrial problems are worth effort investment from university groups • High business/industry value • Allow research publications (not business secret) to advance the knowledge • Challenging problem (does it need highly intellectual university researchers?) • Desirably real man-power investment from both sides • My recent examples • Tencent WeChat [FSE’16 Industry], [ICSE’17 SEIP]: Android app testing/analysis • Exploring collaborations with Baidu, Alibaba, Huawei • Exploring new collaborations with MSRA SA
- 11. Sustained Productive Academia-Industry Collaborations • Careful selection of target problems/projects • Desirable to start with money-free collaborations(?) • If curiosity-driven nature is also from industry (lab) side, watch out. • Each collaboration party needs to bring in something important and unique – win-win situation • High demand of abstraction/generalization skills on the academic collaborators to pursue research upon high-practice-impact work. • Think more about the interest/benefit of the collaborating party • (Long-term) relationship/trust building • Mutual understanding of expected contributions to the collaborations • Balancing research and “engineering” • Focus, commitment, deliverables, funding, …
- 12. Academia-Industry Collaboration • Academia driven • Industry driven
- 13. Academia-Driven: Research Dissemination • Publishing research results technologies there adopted by companies, e.g., ICSE 00 Daikon paper by Ernst et al. Agitar Agitator https://homes.cs.washington.edu/~mernst/pubs/invariants-relevance-icse2000.pdf ASE 04 Rostra paper by Xie et al. Parasoft Jtest improvement http://taoxie.cs.illinois.edu/publications/ase04.pdf PLDI/FSE 05 DART/CUTE papers by Sen et al. MSR SAGE, Pex http://srl.cs.berkeley.edu/~ksen/papers/dart.pdf http://srl.cs.berkeley.edu/~ksen/papers/C159-sen.pdf
- 14. Academia-Driven: Research Commercialization • Commercializing research results in startup tools/products used by companies, e.g., Having a startup ! leading to huge adoption
- 15. Academia-Driven: Tool Community Building • Release open source infrastructures or libraries to engage academic/industry communities to use and contribute, e.g., • MPI/PETSc by Bill Gropp et al. • Charm++ by Laxmikant (Sanjay) Kale et al. • LLVM by Vikram Adve, Chris Lattner, et al. “The openness of the LLVM technology and the quality of its architecture and engineering design are key factors in understanding the success it has had both in academia and industry.”
- 16. Industry-Driven: Infrastructure • Making infrastructure available for academia to build upon, e.g., Nikolai Tillmann, Jonathan de Halleux, and Tao Xie. Transferring an Automated Test Generation Tool to Practice: From Pex to Fakes and Code Digger. In Proceedings of the 29th IEEE/ACM International Conference on Automated Software Engineering (ASE 2014), Experience Papers http://taoxie.cs.illinois.edu/publications/ase14-pexexperiences.pdf http://research.microsoft.com/pex/
- 17. Industry-Driven: Data • Making data available • inside the company (visiting professors, student interns) • to academia Kika Emoji Keyboard http://research.microsoft.com/codehunt/
- 18. Industry Academia “Tensions” •Academia (research recognitions, e.g., papers) vs. Industry (company revenues) •Academia (research innovations) vs. Industry (likely involving engineering efforts) •Academia (long-term/fundamental research or out of box thinking) vs. Industry (short-term research or work) • Industry: problems, infrastructures, data, evaluation testbeds, … • Academia: educating students, …
- 19. Summary: (How) Can A University Group Do It? • Start a startup • but desirable to have right people (e.g., former students) to start • Release free tools/libraries to aim for adoption • but a lot of efforts to be invested on “non-researchy” stuffs • Collaborate with industrial research labs • but many research lab projects may look like univ. projects • Collaborate with industrial product groups • but many problems faced by product groups may not be “researchy”
- 20. Optimizing “Research Return”: Pick a Problem Best for You Your Passion (Interest/Passion) High Impact (Societal Needs/Purpose) Your Strength (Gifts/Potential)Best problems for you Find your passion: If you don’t have to work/study for money, what would you do? Test of impact: If you are given $1M to fund a research project, what would you fund? Find your strength/Avoid your weakness: What are you (not) good at? Find what interests you that you can do well, and is needed by the people Adapted from Slides by ChengXiang Zhai, YY ZHou
- 21. Experience Reports on Successful Tool Transfer • Yingnong Dang, Dongmei Zhang, Song Ge, Ray Huang, Chengyun Chu, and Tao Xie. Transferring Code-Clone Detection and Analysis to Practice. In Proceedings of ICSE 2017, SEIP. http://taoxie.cs.illinois.edu/publications/icse17seip-xiao.pdf • Nikolai Tillmann, Jonathan de Halleux, and Tao Xie. Transferring an Automated Test Generation Tool to Practice: From Pex to Fakes and Code Digger. In Proceedings of ASE 2014, Experience Papers. http://taoxie.cs.illinois.edu/publications/ase14-pexexperiences.pdf • Jian-Guang Lou, Qingwei Lin, Rui Ding, Qiang Fu, Dongmei Zhang, and Tao Xie. Software Analytics for Incident Management of Online Services: An Experience Report. In Proceedings ASE 2013, Experience Paper. http://taoxie.cs.illinois.edu/publications/ase13-sas.pdf • Dongmei Zhang, Shi Han, Yingnong Dang, Jian-Guang Lou, Haidong Zhang, and Tao Xie. Software Analytics in Practice. IEEE Software, Special Issue on the Many Faces of Software Analytics, 2013. http://taoxie.cs.illinois.edu/publications/ieeesoft13-softanalytics.pdf • Yingnong Dang, Dongmei Zhang, Song Ge, Chengyun Chu, Yingjun Qiu, and Tao Xie. XIAO: Tuning Code Clones at Hands of Engineers in Practice. In Proceedings of ACSAC 2012. http://taoxie.cs.illinois.edu/publications/acsac12-xiao.pdf