Grants talk:IEG/Wikintelligence

My comments

I am going to dive in here and just say that I don't think this proposal is worth funding. If you can name just 5 Wikipedians on any project that support your proposal, then I am willing to reconsider. The main problem I have with this proposal after reading it through, is the use of terminology without any concern for the problematic area the proposal addresses, namely machine-readable text for Wikipedia. If you have followed any of the recent internal debates about the (mis)use of templated infoboxes, then you would know that you really need to tiptoe around these issues. As it reads right now, there is no way this proposal is going to get any kind of buy-in from the user community. A somewhat more humble approach is needed, such as calling the project "Wikisuggest" (a device that suggests things that editors could add to improve Wikipedia pages) rather than "Wikintelligence" (a device that implies it is in and of itself more intelligent than any existing editor) is called for. Jane023 (talk) 13:24, 4 October 2013 (UTC)

Hello Jane, thank you for your comment!Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

I saw you comment on Wikidata chat and came to look here for more details. I agree with Jane that you don't seem to have thought this through from the customers point of view. That's like going to a job interview and telling the interviewer how much you would benefit from the job rather than how much you could contribute to the company.

The larger wikipedias have benefited from editors willing to research information, weigh different sources and summarise the information in readable articles. Your project doesn't seem to be able to contribute to any of these tasks.

Wikidata has started to accumulate data in a form that is readable by bots but even the best wikidata pages includes only the bare facts. Turned into prose these facts would barely make a lead paragraph for a wikipedia article. That's ok. That is all wikidata is meant to do.

Just as school students read wikipedia but university students should be writing wikipedia so Artificial Intelligence needs to start by learning to read Wikipedia. Wikidata is designed to help with that. Creating an AIs write wikipedia is a whole different thing, especially considering that there are few actual humans who can do that.

Of course there could be more humans making useful contributions to wikipedia if they had an intelligent sidekick to help them by making suggestions. This would seem to be a more useful niche for your bot. This may even be what you meant to propose here. If so then you need to do a rewrite with some use cases telling us what this bot can do for us! Filceolaire (talk) 15:09, 5 October 2013 (UTC)

Hello Filceolaire, thank you for your comment!Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

Timeline for IEG

Hi Edgar,

Thanks for taking the time to submit this proposal! I note that you're hoping to move forward with the feasibility study in early December, and just wanted to point out that if the IEG committee does recommend that WMF fund this project, our timeline is such that approvals would not happen until after December 5. We'll be announcing those who are selected for a grant on December 16th, and funds would be disbursed soon after that. I mention this so that you can take this into account as your plans move forward with other partners, etc. Best wishes, Siko (WMF) (talk) 18:42, 4 October 2013 (UTC)

Hello Siko, thank you for your comment, for the information and for your good wishes.Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

Eligibility confirmed, round 2 2013

This Individual Engagement Grant proposal is under review!

We've confirmed your proposal is eligible for round 2 review. Please feel free to ask questions here on the talk page and make changes to your proposal as discussions continue during this community comments period.

The committee's formal review for round 2 begins on 23 October 2013, and grants will be announced in December. See the schedule for more details.

Questions? Contact us.

Some questions and comments

Hi, this is based on my first few minutes looking at this proposal so I will likely have more to say. For now,

• What was the reason given by FFG for their willingness to fund this project?

The FFG funded a so-called "Innovationsscheck" (innovation voucher) to check if a research idea of a SME is worth being further pursued. This innovation voucher is worth 5.000 Euro and can be used for buying research from universities or from approved research institutions like OFAI. Because the result of the innovation voucher is, that the idea is worth being further pursued, I am optimistic that the FFG would fund the project, but FFG will decide an 4 December and they would fund only 75% of costs.Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

• Who at the FFG gave this permission and can we talk with them?

Dr. Henrietta Egerth and Dr. Klaus Pseiner are the executive directors who signed the innovation voucher on behalf of FFG. The telephone number of the hotline of FFG for this is +43 5 7755-0. Feel free to contact them or their staff. If you need a copy of the innovation voucher I can mail it to you as a confidential information.Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

• What specific benefits will this proposal have for the Wikimedia community? I am not clear on what the benefits are to editors or readers. I see how phase 3 could be useful but that is a long way from the feasibility study. The benefits of phase 4 and 5 to Wikimedia editors and contributors are not clear to me. Also it is not clear to me that the feasibility study alone is of much benefit to Wikimedia editors, since it is possible that one or more bots in the future may do what is proposed for phase 3 here and I am not sure that the proposed use of resources at this time, in this way, is likely to be a good investment of Wikimedia resources.

the specific benefits for the community are two: first there will be a realistic feasibility study how the goal "Wikintelligence" is reachable with state-of-the-art Artificial Intelligence methods within the next years. This would be another project of course. Second the way to the feasibility study itself and to funding could open a new way of getting money (like Innovation voucher, feasibility study) that is suitable also for other Wikimedia projects. Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

• It is not clear to me how this feasibility study will be conducted. Can you explain your methodology and give a more detailed timeline?

The OFAI will give me an offer which will include:

• problem statement
• method of resolution, state of the art
• planned work, restriction of workload
• project leader, project members
• goal and costs

The feasibility study itself will contain:

• definition of task and problem
• technical specifications with regards to content and specifications for cost-benefit relation
• goal and framework and scope of the feasibility study
• method of resolution and alternatives
• information about experiments
• estimation of chance of realisation and possibilities

Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

• Have you talked with the Wikimedia Legal Department about your proposed use of Wikimedia trademarks? The use of trademarks is taken seriously by the community and WMF. The proposed use seems to imply an endorsement from WMF and the community that I'm not sure is appropriate.

No, I have not talked to them yet. I want to use the trademarks because if Wikimedia gives the grant I think it should be clear that the study was cofinanced by Wikimedia and therefore the logos should be on the report.

Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

Thanks, --Pine^✉ 07:18, 7 October 2013 (UTC)

Thank you very much for the questions! Projekt ANA (talk) 20:58, 11 October 2013 (UTC)

Hi, I have read your responses. I think this is an interesting project but I have a lot of reservations about using Wikimedia funding for it. I think it would be interesting to talk with Dr. Egerth and Dr. Pseiner. Could you ask them if at least one of them would be willing to have a meeting on IRC with me and perhaps other members of the IEG Committee? --Pine^✉ 06:34, 19 October 2013 (UTC)

Hello, today I tried to talk to Dr. Pseiner or Dr. Egerth. Members of the staff (FFG has 264 employees according to their homepage) redirected me to Ms. Mag. Gerlinde Tuscher, head of the public relations department (Tel.: +43 5 7755-6010, email: gerlinde.tuscher (at) ffg.at). She said she would appreciate very much to get an email with the questions that you would like to ask. Dr. Pseiner and Dr. Egerth are quite busy so it could be difficult to arrange a time and date for an IRC, and it depends on the questions who should be involved with that chat. Are the questions about the specific innovation voucher that I could use, or are they about the innovation voucher programme in general, or about the planned feasibility study or again in general about funding of research activities? (In 2012 FFG funded nearly 3.000 new projects with approximately 500 millions Euro, and 5.000 ongoing projects were supported.) - Projekt ANA (talk) 20:30, 22 October 2013 (UTC)

Hi, my questions are mostly about the feasibility study. I would like to hear more about the reasons that the FFG believes that this feasibility study should be funded. Can Ms. Tuscher explain FFG's reasons for wanting to proceed with this feasibility study? --Pine^✉ 06:53, 23 October 2013 (UTC)

Hello Pine, as far as I understand FFG wants to promote research because it has the statutory task doing so, especially for SMEs (small and medium enterprises). So if there is an offer of a university or of an accredited research institute like OFAI, then they are willing to fund a proposal. But, as I described in my proposal under scope and activities, they will decide about it on 4 December, 2013, and first I have to get the offer of OFAI (unfortunately there is a delay for the offer at the moment). Second reason for FFG to give the grant is, that they only give 75 % of the costs, so the SME has to find a sponsor for the other 25 %.

Projekt ANA (talk) 06:51, 24 October 2013 (UTC)

OK, I will take this answer into consideration when I vote on this project. --Pine^✉ 06:46, 25 October 2013 (UTC)

Community Notifications

Hi Dr. Hagenbichler,
As you know, the IEG committee will begin their review of round 2 proposals on 23 October 2013. To expedite proposal review, I'm looking over the community notifications section of each eligible application. It looks like the link to your post on the Wikidata-Project chat was archived. Please re-link to your post's present location. Also, in your engagement targets, you include Commons and Wikipedia (German, English, Simple English, and Spanish). It would be good if you could include any communication you've had with community members on those projects. Please paste links to where you've notified them about your proposal. And please do so soon so that the committee may easily follow up. Thanks!
Best of luck! :)
Anna Koval (WMF) (talk) 23:49, 21 October 2013 (UTC)

Hi Mrs. Koval, thank you very much for your comments, hints and good wishes! I re-linked the notification and included new posts for Commons and Wikipedias (en, simple, de, es). Projekt ANA (talk) 21:49, 22 October 2013 (UTC)

P.S. The committee may also be curious to know if you've contacted Drs. Egerth and Pseiner, as Pine suggested above. Anna Koval (WMF) (talk) 23:54, 21 October 2013 (UTC)

I tried to contact them (see above). Thank you! Projekt ANA (talk) 21:49, 22 October 2013 (UTC)

goals need to be clarified

First time I read this I thought it was unrealistic and out of scope. Second time around I was a bit more favourable to it-identifying if wikidata would be able to be used effectivly as a knowladge base for an open natural language conversation bot could potentially be useful to determine how effevtive of a knowladge base wikidata is and if anything should be changed. But many of the goals here seem super pie in the sky. They need to be trimmed down and made more precise imo (including reasoning as to why X should be included as a goal). I also wouldn't support it unless one of the wikidata devs said its remotely sane. I'm not an ai expert, this could be snake oil for all I know. Bawolff (talk) 21:36, 22 October 2013 (UTC)

Hello Bawolff, thank you very much for your comment! You are exactly right, and that is exactly what the project tries - to make a feasibility study what is realistic and what is unrealistic. That is why the AI experts of the Austrian Research Institute for Artificial Intelligence (OFAI) are involved and they do the work for us, but we Wikimedians can tell them what we need ...

Projekt ANA (talk) 21:58, 22 October 2013 (UTC)

I can't make head or tail of this

You say this is about "using Wikidata as open Artificial Intelligence for improving quality of Wikipedia articles" but you never define (nor do you link to any definition of) "open Artificial Intelligence" nor do you explain, even by example, how it will "improv[e the] quality of Wikipedia articles." Admittedly, I only spent 5 minutes looking at this, but I have a Masters in Computer Science and over 30 years experience in the field, and if I can't make basic sense of it that quickly, it needs to be clearer. - Jmabel (talk) 00:32, 23 October 2013 (UTC)

Hello Jmabel, thanks for the comment. I will try to make it clearer. "Open" in this context just refers to the open source idea, so the project reports and prototypes will be released under a free license like CC-BY-SA or GNU-FDL and so on. Artificial Intelligence is defined as passing the Turing Test as performed in the Loebner contest. These are the Roadmap-subgoals 4 and 5. The examples, how it will "improve the quality of Wikipedia articles are in Roadmap-subgoal 3: "Until 2016 the Wikintelligence project members will develop performance criterions, show cases and use cases within Wikidata and Wikipedia. This will include 1. detecting and defending spam attacks, 2. detecting of inconsistencies between different language versions of an Wikipedia article, 3. checking currentness of data and references of an Wikipedia article and (semi-)automatic updating, 4. (semi-)automatic categorizing of Wikimedia Commons media, 5. (semi-)automatic update of help pages of Wikidata".Projekt ANA (talk) 15:54, 24 October 2013 (UTC)

Jmabel, thanks for commenting and I agree that parts of this request are vague. I am hoping that an explanation from the FFG will give us a better understanding of this grant proposal. --Pine^✉ 06:56, 23 October 2013 (UTC)

As far as I can tell, only roadmap subgoal three is relevant to us: "This will include 1. detecting and defending spam attacks, 2. detecting of inconsistencies between different language versions of an Wikipedia article, 3. checking currentness of data and references of an Wikipedia article and (semi-)automatic updating, 4. (semi-)automatic categorizing of Wikimedia Commons media, 5. (semi-)automatic update of help pages of Wikidata." Of these 5 sub-sub goals, only the first and third are worth funding, as "inconsistencies between different language versions of an Wikipedia article" is not considered a problem, efforts for categorizing on Commons are only useful in combination with Wikidata, and (semi-)automatic update of help pages will probably be frowned upon by the community, as semi-automatic updates of any pages at all are hard to get approval for. Jane023 (talk) 13:01, 24 October 2013 (UTC)

Hello Jane, thank you very much for your comment. I am happy to hear that subgoal 3 with its sub-subgoals 1. and 3. seem to be relative clear and relevant. I will try explain my thoughts to sub-sub-goal 2 "inconsistencies between different language versions": I do not define inconsistency as different styles or different representation of facts, but I want to define it as different representation of facts in different languages, where at least one article is definitly wrong. E.g. each article is right with its references, but in one language there are newer facts and newer references that contradict the older facts of the other articles. Wikintelligence should detect these kinds of inconsistencies. Relating to Sub-subgoal 4: of course you are right, efforts for categorizing on Commons are only useful in combination with Wikidata. The project combines basically and essentially Wikidata with Commons with the use of state-of-the-art Artificial Intelligence algorithms (vision, music, speech, multi-modal data). Relating to Sub-subgoal 5 or the "semi-automatic updating": I am aware of the problems if there is an automatic update of articles in the communities. That is why I defined it as (semi-)automatically, so that it can be used as a tool for the authors, and the authors decide if they will make the update based on the suggestion of Wikintelligence. I will try to explain my thoughts about subgoal 1 and 2 (human-neurophysiological properties together with the virtual robot platform and rule-based clustering data classification techniques for usage on Wikidata within the Pywikipediabot): IMHO these are the necessary steps to reach subgoal 3, based on the concepts of embodied cognition. Subgoal 4 and 5 are surrogates for machine learning and deep understanding of meaning of Wikipedia articles as well as human communication and the dimensions of the world we live in. Can machines think? The Turing Test is the most well known standard for defining Artificial Intelligence since more than 60 years, although it is still under heavy discussion. I assume that Wikintelligence will be able then to contribute to the vision, that every single human being can freely share in the sum of all knowledge. Projekt ANA (talk) 22:26, 25 October 2013 (UTC)

European Network for the Advancement of Artificial Cognitive Systems, Interaction and Robotics

Today I wrote an article in German about the European Network for the Advancement of Artificial Cognitive Systems, Interaction and Robotics. This network would be contacted and asked for collaboration with our project. Projekt ANA (talk) 19:04, 2 November 2013 (UTC)

Aggregated feedback from the committee for Wikintelligence

Scoring criteria (see the rubric for background)	Score 1=weakest 5=strongest
Potential for impact
(A) The project fits with the Wikimedia movement's strategic priorities	3
(B) The project has the potential to lead to significant online impact.	3
(C) The impact of the project can be sustained after the grant ends.	2
(D) The project has potential to be scaled or adapted for other languages or projects.	3
Ability to execute
(E) The project has demonstrated interest from a community it aims to serve.	2
(F) The project can be completed as scoped within 6 months with the requested funds.	2
(G) The budget is reasonable and an efficient use of funds.	3
(H) The individual(s) proposing the project have the required skills and experience needed to complete it.	3
Fostering innovation and learning
(I) The project has innovative potential to add new strategies and knowledge for solving important issues in the movement.	3.5
(J) The risk involved in the project's size and approach is appropriately balanced with its potential gain in terms of impact.	2.5
(K) The proposed measures of success are useful for evaluating whether or not the project was successful.	3
(L) The project supports or grows the diversity of the Wikimedia movement.	2.5
Comments from the committee: Overall idea is interesting, but the proposal could have been more specific. A very theoretical approach with only indirect expected benefit to our community. This futuristic idea may be useful to Wikimedia further down the road. We have some concerns about what would be doable in the time frame planned and for the money planned, and we would like to see more direct impact for this cost. Community skepticism and lack of community endorsements are a concern. "Semi-automatic" edits may be problematic for some community members. Some concerns about the project team lacking necessary skills or background.

Thank you for submitting this proposal. The committee is now deliberating based on these scoring results.

Funding decisions will be announced by December 16. — ΛΧΣ²¹ 00:19, 14 November 2013 (UTC)

Status update

This project has not been selected for an Individual Engagement Grant at this time.

We love that you took the chance to creatively improve the Wikimedia movement. The committee has reviewed this proposal and not recommended it for funding, but we hope you'll continue to engage in the program. Please drop by the IdeaLab to share and refine future ideas!

Next steps:

1. Review the feedback provided on your proposal and to ask for any clarifications you need using this talk page.
2. Visit the IdeaLab to continue developing this idea and share any new ideas you may have.
3. To reapply with this project in the future, please make updates based on the feedback provided in this round before resubmitting it for review in a new round.
4. Check the schedule for the next open call to submit proposals - we look forward to helping you apply for a grant in a future round.

Questions? Contact us.

Digital library

If the team could get free access to a digital library, I think you could try using AI to find reference for a lot of articles. The robot don't need (in the beginning) update references, it can just point possible references (it could give points to every possible reference for a sentence) and notify users that have edited the article being analyzed, or users that belong to a specific group of user.

I think that would be an easy, and important, starting point.

Thanks for the initiative, I'm not a professional in the area, but I like the use of technology to improve our lives! Guiwp (talk) 01:44, 23 December 2013 (UTC)