"TextCollab team on OpenAI Whisper, GPT3, Codex & DALL-E 2 Hackathon Hackathon"

Team Idea

Alternative name (BitVote.net) full concept here: http://bitvote.net/?page_id=41 Relevant description / summary Alternative name (BitVote.net) entire concept here: http://bitvote.net/?page_id=41 Relevant description/summary textCollab is likely going to be stage 1 to 3 from the whole BitVote idea above. BitVote will start to operate at the discussion and proposal creation level. Off-setting the discussion to more internet platforms (such as Reddit, in the future). As step one in the process, textCollab/BitVote users may pick up a popular question or issue in their respective community. In the next step, the users start proposals sessions on this platform to address the question or any of the available questions. Anyone can start a question or issue on BitVote, similar to sites like Quora. If the case matches the keywords or content of another already in the system, the user is encouraged to review the previous one. He can still go ahead and create a similar proposal or text document, but the algorithm will still group it thematically based on content and keywords. As soon as that is done, anyone that is registered with the system can see new issues by region, type, language, or other keywords. A brand new issue or question will still need to have proposals in it. One can start writing a proposal for the question at hand. One of the essential central features of BitVote, is the editor, and this will be the textCollab project. As the user is writing the algorithm is dynamically scanning the content. Searching through any sources (Wikipedia, internet archives, academic journals, meta-searches in search engines), and presenting the results as a part of the screen (for computers as devices; a secondary tab with notices for mobile devices). One additional source will be the textCollab database, scanning for similarities between the current and previous proposals already in the system. From the thematic content analysis, the user can see relevant papers and sources that can quickly inform him to improve his proposal or use them as sources. Suppose the proposal is very high in similarity with another. In that case, the user will be prompted to join in support of that proposal unless there are any additions or sub-versions to that proposal. The aim is to reduce the total amount of proposals and foster synthesis. Thus, creating more complete proposals from many authors in the system. This stage can take weeks, depending on the configuration of the person who is running this proposal instance has chosen. Furthermore, there may be competing proposal chains in the system with different configurations. If both are set to public view, both can benefit from existing content and work in the textCollab archive. On the other hand, there are cases where a BitVote process will be started by a corporate entity or organization where only select groups of people will have electronic access keys to see the full content (or any configuration of the above). In the case of competing proposals, one may arrive at a solution using different time periods, different groups of people and different implementer(s) or auditor(s) of the voted solution. The content ranking should be done continuously but iterate until after the Proposal period. The rank value of every proposal (or synthesis of proposals, for that matter, for larger groups) is based on some of the following: Support by primary academic sources. (Max 0.5) Support by other literature. (Max 0.2) Education or specialization of the author (thematic match). (Max 0.1) Thematic match to the question addressed. (Max 0.1) User ranking and ratings. (Max 0.1) Potentially other factors. The voting stage is generally a simpler process but depending on the whole session or BitVote process configuration one can configure a vote with different rules appropriate for the occasion. And what should happen if a satisfactory result is not reached? Return to Proposal session for a set amount of time to foster more formal discussion, or re-iterate the Vote session until there is a clear result. Assuming there is a clear winner, this part of the process is done, now depending on the BitVote process configuration there can be a vote for who will be the implementer or a public competition for potential implementers to compete. The next stage can be assigned to an auditing authority that will assess and report back in the system how the implementation of the proposal went. All this stays in archival to be used continuously in the system and also for historical record. If a great proposal and implementation is reached that will be featured and highlighted to see what has been learned, but the same really for a failed proposal or implementation. We need such methods to be able to cope with the amount of information we will have to deal with in a large scale application of meritocratic democracy today – that means very big populations and thus countless submissions of proposals to dig through. That would be impossible without a smart collaboration design. The intent is to let the best ideas come forth, be expressed and compete. Choose democratically from those, and then implement and assess.
Flibbert_debola
Adebola Rabiu
Flibbert_debola

Data scientist

Mwandium1999
Chibuyu Mwando
Mwandium1999

Student

osamatech786
Muhammad Osama Ahmed
osamatech786

Data Scientist

Phoebusg
Phoebus Giannopoulos
Phoebusg

IT

__Ashi__
Ashish Abraham
__Ashi__

Software Engineer

Krishna7604
Krishnavardhan Ponaka
Krishna7604

Student

Submission

Team Leader hasn't made a submission yet