Jonny: Undo revision 444 by Jonny (talk)

2022-10-26T21:06:11Z

Undo revision 444 by Jonny (talk)

← Older revision		Revision as of 21:06, 26 October 2022
Line 1:		Line 1:
	{{Participant		{{Participant
	\|Timezone=America/~~New York~~ (~~GMT−05~~:00/~~GMT−04~~:00)		\|Timezone=America/Argentina/Buenos Aires (GMT−03:00/GMT−03:00)
	\|Affiliation=~~Knowledge Futures~~		\|Affiliation=Minerva University
	\|Projects=~~PubPub, WikiFunctions, iNat, Geograph.org.uk~~		\|Projects=Abstract Poetry
	\|Table Assignment=Table 4		\|Table Assignment=Table 4
	}}		}}

Jonny at 20:27, 26 October 2022

2022-10-26T20:27:08Z

← Older revision		Revision as of 20:27, 26 October 2022
Line 1:		Line 1:
	{{Participant		{{Participant
	\|Timezone=America/~~Argentina/Buenos Aires~~ (~~GMT−03~~:00/~~GMT−03~~:00)		\|Timezone=America/New York (GMT−05:00/GMT−04:00)
	\|Affiliation=~~Minerva University~~		\|Affiliation=Knowledge Futures
	\|Projects=~~Abstract Poetry~~		\|Projects=PubPub, WikiFunctions, iNat, Geograph.org.uk
			\|Table Assignment=Table 4
	}}		}}
	{{Workshop Submission		{{Workshop Submission

Jonny at 06:13, 21 October 2022

2022-10-21T06:13:50Z

← Older revision		Revision as of 06:13, 21 October 2022
Line 24:		Line 24:

	Our Team Page: https://curlprojects.com/		Our Team Page: https://curlprojects.com/
			\|Organizer Topics=Discovery, Search, Contextualization, NLP, Bibliometrics
	}}		}}

Jonny: Created page with "{{Participant |Timezone=America/Argentina/Buenos Aires (GMT−03:00/GMT−03:00) |Affiliation=Minerva University |Projects=Abstract Poetry }} {{Workshop Submission |Interest=Our project, called Abstract poetry, is a semantic search engine which represents the contextual relationships between academic papers. With this first prototype, we aimed to push search from a frustrating, biased-by-popularity, and always incomplete process to one that was simple, unbiased by design..."

2022-10-20T04:30:57Z

Created page with "{{Participant |Timezone=America/Argentina/Buenos Aires (GMT−03:00/GMT−03:00) |Affiliation=Minerva University |Projects=Abstract Poetry }} {{Workshop Submission |Interest=Our project, called Abstract poetry, is a semantic search engine which represents the contextual relationships between academic papers. With this first prototype, we aimed to push search from a frustrating, biased-by-popularity, and always incomplete process to one that was simple, unbiased by design..."

New page

{{Participant
|Timezone=America/Argentina/Buenos Aires (GMT−03:00/GMT−03:00)
|Affiliation=Minerva University
|Projects=Abstract Poetry
}}
{{Workshop Submission
|Interest=Our project, called Abstract poetry, is a semantic search engine which represents the contextual relationships between academic papers. With this first prototype, we aimed to push search from a frustrating, biased-by-popularity, and always incomplete process to one that was simple, unbiased by design, and could both represent high level patterns in academic literature while also exhaustively returning all relevant results.

But we’ve currently hit a couple of roadblocks that we would like to work out with people in the workshop. First, Abstract Poetry currently can’t understand papers at ‘multiple’ levels. Our algorithm understands them as similar, but not as related in terms of methodology vs related in terms of topic. This limits the search to returning general patterns and makes it inconsistent at returning exhaustive results for literature reviews. Second, we can’t return an accurate number of topics related to a search. Instead we currently hard code an average number of topics, this can sometimes force papers together that aren’t related limiting the usefulness of the discovered patterns for researchers. Finally, because our search doesn’t understand different ‘dimensions’ of search, sometimes it can fixate on patterns the researcher doesn’t care about.

What we would like to do is work with other people who are deeply knowledgeable about the most relevant patterns, relational knowledge technologies, and researcher workflows to figure out both technical and design solutions to the difficulties mentioned above so that Abstract Poetry can allow researchers to more effectively navigate the academic ecosystem!
|Frame=Tool-builder
|Materials=To provide a search that can more reliably return relevant results across even bigger stores of knowledge that we have now, we have created two working projects -

The first is a search engine. Abstract Poetry Search uses semantic embeddings to understand papers and their meaning and groups them based on similarity. It then asks researchers to react to papers letting the algorithm know which patterns and topics the researcher is most interested in. With this search, results are already grouped into the relevant fields that researchers want to explore. When a researcher wants to exploit a particular group of papers for all the related literature to their question, they can do so by simply continuing to leave positive impressions on papers in that group until the group stops returning relevant results. In theory, this could allow for an exhaustive search which returns all relevant papers in a specific field without returning to the search bar to change your query. Currently, however, our database is limited to PLOS, and the technical limitations above limit us from being fully exhaustive.

The second is a semantic bibliography. Abstract Poetry Bibliography takes in the DOI of a paper that a researcher would like to explore and returns all the references of that paper in semantic groups. This provides the researcher a high level view of the fields of research that the paper is pulling from and can allow the researcher to more quickly find the relevant research papers from a paper bibliography instead of surfing back and forth through the text and the references section.

Abstract Poetry Search: https://abstract-poetry.fly.dev/search (select the take a tour button below the search bar to learn how to interact with the interface)

Abstract Poetry Bibliography: https://abstract-poetry.fly.dev/bibliography

Abstract Poetry Landing Page: https://readymag.com/u192065531/3832227/

Our Team Page: https://curlprojects.com/
}}

Elianna DeSota - Revision history

Jonny: Undo revision 444 by Jonny (talk)

Jonny at 20:27, 26 October 2022

Jonny at 06:13, 21 October 2022