Skip to main content
SpringerLink
Log in

Context-based counselor agent for software development ecosystem

  • Published:
Computing Aims and scope Submit manuscript

Abstract

Counseling for information technology (IT) personnel lies at the intersection between the software development ecosystem where IT employees collaborate professionally and the social ecosystem where they communicate with each other to share the success or handle the failure of software development. Today, counseling has become a major issue in the IT industry, since the success rate of IT system development projects is as low as 30 %, and more than 60 % of IT professionals suffer from anxiety or other emotional problems. This paper describes a conversational agent aiming to replace human counselors assisting IT personnel in software development ecosystems toward future deployment to social ecosystems. Utilizing IT domain ontology knowledge, our agent automatically adapts the vocabulary used in its responses according to the context and to the current phase of the conversation. Using context-based reflection support knowledge, the agent generates its response consisting of (1) chatterbot-like mirroring/rewording for context sharing and (2) newly proposed context-respectful mechanism of prompts for “context narrowing/digging” to help a client discover problems and become aware of their solutions via deep reflections of IT personnel undergoing counseling. Knowledge focusing on a single domain, such as IT counseling domain, and context-based/context-respectful reflection allow our counseling agent to work properly without having to acquire and manage a huge amount of knowledge. Experimental results show that clients interact with our agent on average two times longer than they do with ELIZA-style conversational agents; also, a questionnaire-based validation has shown the average value of questionnaire’s result was “agree” side for our agent, but “disagree” side for ELIZA-style conversational agents. Therefore, the user acceptance level of our agent is much higher than that of conventional chatterbots.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2
Fig. 3
Fig. 4
Fig. 5
Fig. 6

Similar content being viewed by others

Notes

  1. It is interesting to remark that only in the year 2000s, within the Loebner competition, conversational agents’ performance managed to get back to the PARRY levels of nearly 30 years before, in spite of all the theoretical activity carried out since then.

  2. Job categories explained in ITSS Career Framework: marketing, sales, consultant, IT architect, project management, IT specialist, software development, customer service, IT service management, education.

  3. This access shuffling technique reduces the influence of the recruitment (e.g., in terms of the participant skills and previous knowledge) on the experiment outcome [40].

  4. Somewhat surprisingly, few of our IT students seem to have heard about ELIZA.

  5. Normal distribution was confirmed about TRUST in CA(p = 0.200 \(>\) 0.05)/EL(p = 0.200 \(>\) 0.05) and SELF-AWARENESS in CA(p = 0.200 \(>\) 0.05)/EL(p = 0.200 \(>\) 0.05) via Kolmogorov–Smirnov analysis on the data of 15 persons.

  6. The actual conversation used our Japanese language module. Here, we provide an English translation for the sake of understandability.

References

  1. Bosch J (2009) From software product lines to software ecosystems. In: Proceedings of SPLC, 13th international software product line conference, San Francisco, pp 111–119

  2. Bosch J, Bosch-Sijtsema PM (2010) From integration to composition: on the impact of software product lines: global development and ecosystems. J Syst Softw 83(1):67–76

    Article  Google Scholar 

  3. Yaguchi R, Yoshida Y (2008) Project success rate: 31.1 %. Nikkei Computer, Japan, pp 36–53

    Google Scholar 

  4. Ministry of Health, Labour and Welfar (2008) General condition of mental health caring and smoking countermeasure consequence, Japan. http://www.mhlw.go.jp/toukei/itiran/roudou/saigai/anzen/kenkou07/index.html. Accessed 19 Jan 2013

  5. Shinozaki T (2008) Active listening power. NTTDATA INTELLILINK, Japan

    Google Scholar 

  6. Kosh MS, Kerzner H (1984) Stress and burnout in project management. Project Management Institute Seminor/Symposium, Penn: PMI, Philadelphia, pp 125–132

    Google Scholar 

  7. Jaaksi A (2007) Experiences on product development with open source. Open source development, adoption and innovation, pp 85–96

  8. Weizenbaum J (1976) Computer power. Saimaru Shuppan, Japan (in Japanese, translated by Akiba T)

    Google Scholar 

  9. Weizenbaum J (1966) ELIZA—a computer program for the study of natural language communication between man and machine. Commun ACM 9(1):36–45

    Article  Google Scholar 

  10. Floridi L, Taddeo M, Turilli M (2009) Turing’s imitation game: still an impossible challenge for all machines and some judges—an evaluation of the 2008 Loebner contest. Minds and machines. Springer, New York

    Google Scholar 

  11. Biswas G, Roscoe R, Jeong H, Sulcer B (2009) Promoting self-regulated learning skills in agent-based learning environments. In: Proceedings of the 17th international conference on, computers in education, pp 67–74

  12. van de Pol J, Volman M, Beishuizen J (2010) Scaffolding in teacher–student interaction: a decade of research. Educ Psychol Rev 22:271–296 (Springer)

    Article  Google Scholar 

  13. Asay TP, Lambert MJ (1999) The empirical case for the common factors in therapy: quantitative findings. The heart and soul of change, pp 23–55.

  14. Schon DA (1983) The reflective practitioner: how professional think in action. Yurumi Shuppan, Japan (in Japanese, translated by Satou M, Akita K)

    Google Scholar 

  15. Rogers C (1966) The complete works of Rogers vol. 3 psychotherapy. Iwasaki Gakujyutsu Shuppansya, Japan (in Japanese, translated by Fujio T)

    Google Scholar 

  16. Gonzalez AJ, Tsuruta S, Sakurai Y, Nguyen J, Takada K, Uchida K (2010) Using contexts to supervise a collaborative process. Artificial intelligence for engineering design, analysis and manufacturing (AIEDAM). Cambridge University Press, Cambridge

    Google Scholar 

  17. Hung V, Gonzalez A, DeMara R (2009) Towards a context-based dialog management layer for expert systems. IEEE Conference Publications, New York, pp 60–65

    Google Scholar 

  18. Parkinson R, Colby M, Faught W (1976) Conversational language comprehension using integrated pattern-matching and parsing. Artif Intell 9(2):111–134

    Article  Google Scholar 

  19. Thomas E, Chamberlain W (1984) Racter. Computer software, Mindscape

  20. Deryugina OV (2010) Chatterbots. Sci Tech Inf Process 37(2):143–147

    Article  Google Scholar 

  21. Ball G, Breese J (2000) Emotion and personality in a conversational agent. Embodied conversational agents, pp 189–219.

  22. De Pietro O, De Rose M, Frontera G (2005) Automatic update of AIML knowledge base in e-learning environment. In: Proceedings of computers and advanced technology in education, Oranjestad, pp 29–31

  23. Kopp S, Gesellensetter L, Kramer N, Wachsmuth I et al (2005) A conversational agent as museum guide: design and evaluation of a real-world application. In: Panayiotopoulos (ed) Intelligent virtual agents, vol 3661. LNAI, New York, pp 329–343

    Chapter  Google Scholar 

  24. Robinson S, Traum D, Ittycheriah M, Henderer J (2008) What would you ask a conversational agent? Observations of human-agent dialogues in a museum setting. In: Proceedings of the sixth international conference on language resources and evaluation (LREC’08), pp 1125–1130

  25. Sammut C (2001) Managing context in a conversational agent. Mach Intell 18:189–202

    Google Scholar 

  26. Justine C, Stone M, Yan H (2000) Coordination and context-dependence in the generation of embodied conversation. In: Proceedings of the first international conference on natural language generation, volume 14. Association for computational linguistics, pp 171–178

  27. Wallace RS (2003) The elements of AIML style. Alice AI Foundation

  28. Kudo T, Yamamoto K, Matsumoto Y (2004) Applying conditional random fields to Japanese morphological analysis. In: Proceedings of the 2004 conference on empirical methods in natural language processing (EMNLP-2004), pp 230–237. http://mecab.googlecode.com/svn/trunk/mecab/doc/index.html

  29. Dernoncourt F (2012) Designing an intelligent dialogue system for serious games. http://ai-faq.com. Accessed 1 Feb 2013

  30. Tarau P, Figa E (2004) Knowledge-based conversational agents and virtual storytelling. In: Proceedings of the 2004 ACM symposium on applied computing (SAC 04), ACM Press, New York, pp 39–44

  31. Employment Security Bureau, Ministry of Health, Labour and Welfare (2011) Welfare Labor Department of Ed. Occupational classification: special, technical profession (revised 2011), Japan. https://www.hellowork.go.jp/info/mhlw_job_dictionary_02.html. Accessed 19 Jan 2013

  32. INFORMATION-TECHNOLOGY PROMOTION AGENCY (IPA) (2008) Skill standards for IT professionals V3 English edition release 1.0. Ministry of Economy, Trade and Industry, JAPAN. http://www.ipa.go.jp/english/humandev/forth_download.html. Accessed 19 Jan 2013

  33. Isbister K, Doyle P (2005) The blind men and the elephant revisited. In: Ruttkay Z, Pelachaud C (eds) From brows to trust: evaluating embodied conversational agents: human–computer interaction series. Volume 7, Chap. 1. Springer, Netherlands, pp 3–26

    Google Scholar 

  34. Van der Heijden H (2004) User acceptance of Hedonic information systems. MIS quarterly, pp 695–704

  35. DeMara RF et al (2008) Towards interactive training with an avatar-based human–computer interface. Interserv/Ind Train Simul Educ Conf (I/ITSEC) 8054:1–10

    Google Scholar 

  36. Hung V, Elvir M, Gonzalez A, DeMara R (2009) Towards a method for evaluating naturalness in conversational dialog systems. In: Proceedings of the 2009 IEEE international conference on systems: man and cybernetics, pp 1236–1241

  37. Fukushima O, Tsuchida T, Mori M, Matsumoto C, Suzuki A (2007) The effects of one-to-one interview method, group method, and assumed writing in counselor training program. Mejiro J Psychol 3:63–75

    Google Scholar 

  38. Yeh CJ, Chang T, Kowalewska-Spelliscy D, Drost C, Srivastava D, Chiang L (2012) Evaluation of a web-based peer discussion group for counselor trainees. Int J Psychol Behav Sci 2(6):214–225

    Article  Google Scholar 

  39. Saygin AP, Cicekli I, Akman V (2000) Turing test: 50 years late. Minds Mach 10:463–518

    Article  Google Scholar 

  40. Likert R (1932) A technique for the measurement of attitudes. Arch Psychol 22(140):1–55

    Google Scholar 

  41. Ministry of Internal Affairs and Communications Statistics Bureau (2010) Employed persons by industry, status in employment (employee by number of persons engaged in enterprise). http://www.e-stat.go.jp/SG1/estat/ListE.do?lid=000001105914. Accessed 19 Jan 2013

  42. Traum D, Roque A, Leuski A, Georgiou P, Gerten J, Martinovski B, Narayanan S, Robinson S, Vaswani A (2007) A virtual human for tactical questioning. In: Proceedings of the 8th SIGdial workshop on discourse and dialogue, Antwerp, pp 71–74

Download references

Acknowledgments

We gratefully acknowledge KAKENHI (24700214) support by Japanese Society for Promotion of Science. We also deeply thank researchers and students of Distributed Intelligent Systems Lab, Tokyo Denki University, for their dedicated help with modifying our current system and conducting evaluation tests.

Author information

Authors and Affiliations

  1. Graduate School of Advanced Science and Technology, Tokyo Denki University, 2-1200 Muzai Gakuendai, Inzai, Chiba, 270-1382, Japan

    Tetsuo Shinozaki

  2. School of Information Environment, Tokyo Denki University, 2-1200 Muzai Gakuendai, Inzai, Chiba, 270-1382, Japan

    Yukiko Yamamoto & Setsuo Tsuruta

Authors
  1. Tetsuo Shinozaki
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Yukiko Yamamoto
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Setsuo Tsuruta
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Tetsuo Shinozaki.

Appendix: a worked-out example

Appendix: a worked-out example

We shall now discuss a worked-out example of a conversation between our agent and a client.Footnote 6 In the problem-discovery phase, some experiences and accomplishments of the counselor are introduced as a preparation to share contexts in the counseling dialog or conversation, as shown in Fig. 7. In the figure, the acronym CA indicates our counseling agent. Afterwards, the occupation of the client is asked by a prompt to narrow down the client’s situation or conversational context to three categories (IT worker, BSc student, BA student in this example: see Table 5).

Fig. 7
figure 7

Example of greetings and dialogs to focus on or narrow the client’s occupation

Full size image

Then, the conversation continues: in order to build up client’s trust in CA and to promote reflection of client, the client’s accomplishments are asked by prompt and mentioned in the response text.

Figure 8 shows the part of the conversation where client’s trust in CA by highlighting the client’s accomplishments. Questions raised in this dialogue are transformed into responses to the client.

Fig. 8
figure 8

Dialogs to organize the client’s experience and accomplishment

Full size image

The dialogs to ask the client’s experience and accomplishment comply to counseling domain knowledge chunks or patterns selected based on the conversational context (occupation in this case) such as IT worker, BSc student, etc. In case of this BSc student example shown in Fig. 8, three questions are consecutively asked: what did the client accomplish during her high school days, what did she experience and what was especially noteworthy in the experience.

Little by little, the problem that the client is suffering from emerges and is put into words. Then it is classified according to six pre-defined problems (client’s sufferings or problem categories) corresponding to problems that likely to happen at campus or IT workplace: for example, career improvement, school environment, inter-personal relationships (see Table 5). Figure 9 shows the part of the conversation where the client’s suffering is classified according to the above mentioned six problem categories.

Fig. 9
figure 9

Dialogs to clarify and discover client’s problem

Full size image

At the end of problem-discovery phase, the problem the client is suffering from is classified further in detail according to a counseling domain knowledge chunk selected by the current context (career improvement in this example) namely a pre-defined set of ten fine-grained (sub) problems (see Table 5). Consequently, the client’s problem is discovered using the above mentioned procedure or such IT counseling domain knowledge as well as the appropriate context such as the client’s problem for problem-solving phase is set.

In the problem-solving phase, as the conversation on the client problem goes on, related keywords are matched to generate responses aiming to deepen the client’s reflection (Fig. 10). The italic parts in Fig. 10 show how our agent promotes the reflection, i.e. by using reflection-deepening prompts such as “tell me more in detail” or “tell me a little more” or by waiting for client’s input for several minutes are repeated after rephrasing what the client has just said.

Fig. 10
figure 10

Example of dialogs to dig problems/contexts towards solution awareness

Full size image

Rights and permissions

Reprints and permissions

About this article

Cite this article

Shinozaki, T., Yamamoto, Y. & Tsuruta, S. Context-based counselor agent for software development ecosystem. Computing 97, 3–28 (2015). https://doi.org/10.1007/s00607-013-0352-y

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00607-013-0352-y

Keywords

Mathematics Subject Classification (2000)

Search

Navigation