My latest discoveries on library chatbots: Pixel

This week, I learned through Michael Sauers from #NCompassLive Tech Talk about Pixel, the chatbot at the University of Nebraska– Lincoln Library. You can meet Pixel, the reference library chatbot at http://pixel.unl.edu/ and/or view the YouTube presentation on Pixel:





While listening to the presentation, I learned a couple of new things myself. The presenter DeAnn Allison raises the point that chatbots can flatten one's website, and reduce users from having to navigate to get information. Instead, users can just ask a question and the chatbot could provide the information or even the hyperlink to enable the user to get where they want to go, without having to click through several links. Lorna Dawes, the co-presenter, also made the point that chatbots are very useful for library websites that are very dense. As she made this point, I immediately thought of the  Library of Congress website, which I believe is very dense, yet full of useful information. If there is any library that should get one, in my opinion it is the Library of Congress.

The presentation also gave me an opportunity to reacquaint myself with Artificial Intelligence Markup Language (AIML), the language behind the open source chat bot movement. I also learned about Program-O (Open Source PHP MySQL AIML Chatbot software) http://blog.program-o.com/ as another open source software for creating one's chatbots.


Another thing that I discovered or realized during the presentation was the importance of taxonomies or ontologies in developing chatbots. This I realised as DeAnn Allison discussed how she used Library of Congress Subject headings to inform the AIML of a library chatbot. Allison (2012) also has an article on chatbots in the library if you want to check it out.

Reference:
Allison, D.  (2012). Chatbots in the library: Is it time? Library Hi Tech, (30) 1: 95 - 107.

DOI:	10.1108/07378831211213238

Prototypes of virtual agents that provide business information

While browsing Chatbots.org, I have discovered prototypes of virtual conversational agents that do provide business (or company) information online. They are available on this online directory: http://www.chatbots.org/platform/facebook/

I have even discovered pre-existing virtual agents that offer legal and financial information online: http://www.chatbots.org/industry/finance_legal/

Even though I have not evaluated these agents as yet, (which might make for a useful future project), I believe that these discoveries support my perspective that conversational agents can be used to convey business information. It is just a matter of deciding what knowledgebase or database of business information or even company information that one wants a conversational agent to store.


Hence, I see that business librarians can perhaps adapt the conversational agent towards providing company information. Already in business information there is what we commonly call the company profile (See Abels & Klein, 2008, p.42 for definition of company profiles) . Or industry profile for that matter. Hence, a conversational agent could be designed to answer questions about what company produces what products and services. Information provided could also include whether the company is publicly traded or private. The agent could also be designed to indicate the ownership structure of the company.

Reference:

Abels E. G. &  Klein, D. P. (2008). Business information: Needs and strategies. Bingley, UK : Academic Press/Emerald.

Thesis Idea Update: Providing Access to Entrepreneur Blogs through Conversational Agents?

I have been working for the entire summer on two guided reading papers, after which I will prepare for comprehensive examinations, all things going well.

I just wanted to post an update on my thoughts as to a possible direction for my thesis. As some of you know, I have an interest in the use of conversational type agents for providing online information to entrepreneurs. As it stands today, I am not sure if I have the time to develop a prototype adequate enough to be tested for dissertation purposes. Hence, I am considering to just include agents as a chapter in the thesis, perhaps proposing that agents be considered for future in connecting business persons to stories shared on the Web.

 

I guess the larger vision is to create a system for sharing business information, that captures people's business stories in whatever format, and make that accessible to users. This could be through conversational interaction with virtual agents or otherwise (which may involve multimedia information retrieval). 

 

Since no such system exist, but would have to be created, I guess I have limited options, especially constrained by time. I could undertake a project where I would collect the online stories shared by  Jamaican/Canadian entrepreneurs (or both) in whatever online format (social media or regularly Web), and then index and categorise them into motifs. These categories and motifs I would input into multiple conversational agents: the trickster personality for stories that feature wit and intelligence to outsmarting bureaucratic obstacles; the victim personality for stories that deal with tragedy and the hero personality: that deals with stories about overcoming odds and attaining success.

Based on Technorati  2011 report, I know that a large percentage of blogs are created by entrepreneurs who are sharing their stories. I also know that based on a Pew Internet Research report, bloggers are seen as storytellers (Lenhart & Fox, 2006). I see the possibility of indexing and categorising these stories to be the source materials for developing the agents. However, I also believe I could provide access to these entrepreneur blogs or stories  without the use of conversational agents, as I can use a blog or content management system to index and organise a kind of directory to entrepreneur blogs and stories on the Web.

 

Possible research questions thus far for such a project:

 

1. What knowledge is being shared via social media by entrepreneurs' storytelling?
2. Are current motifs and categories of organisational storytelling literature adequate for indexing and categorising entrepreneurs' storytelling via social media?
3. To what extent would making entrepreneur stories accessible via the Web be useful?

 

If I was able to create an adequate agent, I would add this question for testing:

 

4. Would such stories be more accessible if users interacted with virtual agents rather than browsing and keyword searching?

References;

Lenhart, A., & Fox, S. (2006). Bloggers: A portrait of the internet’s new storytellers. Pew Internet & American Life Project. Retrieved from http://www.pewinternet.org/~/media//Files/Reports/2006/PIP%20Bloggers%20Report%20July%2019%202006.pdf.pdf 

Technorati. (2011). State of the blogosphere 2011: Introduction and methodology. Retrieved 6/13/2012, 2012, from http://technorati.com/social-media/article/state-of-the-blogosphere-2011-introduction/

Trickster folkloric characters and my evolving thesis idea

I read an excerpt that I find interesting especially as I consider folkloric characters such as Anansi, Sly Mongoose and Doctorbird as possible options for the AI conversational agent.

"Trickster has particular purposes within cultural stories as educator and teacher. The trickster typically teaches by making mistakes so listeners can hear the stories and learn from the mistakes so they do not have to make them themselves. Trickster typically has some human and some superhuman characteristics and possibly some animal characteristics as well that allow him to get into situations that otherwise would not be possible thus allowing stories to consider complex cultural, emotional, physical, mental, and spiritual domains." (Iseke-Barnes 2009)

Interesting ideas. I never saw it that way. I saw Anasi stories as teaching us the qualities of deception and deceiving other people which I felt ingrained corruption and 'conmanship' in the African diaspora.

My discovery that Native Canadians (or the Indigenous people of Canada) also have trickster stories also makes my thesis idea more relevant to not only Jamaica, but to Canada as well. Hopefully I can establish that AI conversational agents based on one's cultural stories may be effective in the online world in transferring knowledge and information, and in helping people make sense of their activities or world.



Reference:

Iseke-Barnes, J. (2009). Unsettling fictions: Disrupting popular discourses and trickster tales in books for children. Journal of the Canadian Association for Curriculum Studies,

Brief update on my immature thesis idea

Just providing a brief update on my messy thesis concept. Right now it has evolved into the idea of using folkloric artificial intelligent conversational agents for transmitting business information for micro-enterprise owners.

 For the summer, I am hoping to look at literature on folklore being used to convey business information and/or use in online information services. By folklore - I am narrowing down to the use of folkloric storytelling and proverbs as well as the role of natural conversation in using these methods to convey information to micro-enterprise owners.

Also looking for theory to support this perspective. Thoughts, suggestions, or literature welcomed.