Proceedings of the sixth international workshop on Exploiting semantic annotations in information retrieval
Bibliographic record
Abstract
These proceedings contain the contributed papers of the Sixth Workshop on Exploiting Semantic Annotations in Information Retrieval (ESAIR 2013), held at CIKM 2013 in San Francisco, on October 28, 2013. After successful workshops at ECIR'08 in Glasgow, WSDM'09 in Barcelona, CIKM'10 in Toronto, CIKM'11 in Glasgow, and CIKM'12 at Maui, this year's workshop will focus on the need to include large-scale knowledge resources and on annotations beyond the topical dimension. There is a need to include the currently emerging knowledge resources (such as DBpedia, Freebase) as underlying semantic model giving access to an unprecedented scope and detail of factual information. There is also a need to include annotations beyond the topical dimension (think of sentiment, reading level, prerequisite level, etc) that contain vital cues for matching the specific needs and profile of the searcher at hand. ESAIR'13 will be a real workshop where researchers from these different disciplines will work together to identify natural use cases, barriers to success, and work on ways of addressing them: Application/Use Case: What are use cases that make obvious the need for semantic annotation of information? What tasks cannot be solved by document retrieval using the traditional bagof- words? What is keeping searchers from exploring these powerful search requests? What impact has the web of data with more and more information in preprocessed form? Annotations: What types of annotation are available? Are there crucial differences between author-, software-, user-, and machine-generated annotations? Do we annotate types/classes/categories (person) or instances (Albert Einstein)? How similar or different are linked data and annotated text? What are the limitations of the current annotations schemes, and how to overcome them? Rich Context: Do we annotate text? Or also search requests and interactions, and their broader context? Besides personalization and geo-positional information, mobiles have a wide and growing range of locational, mechanical and even biometrical sensor data available to them. Can kick-start the query by inferring task and situational context? (Un)certainty: How should we interpret the annotations? Can we reliably link textual annotations to known entity catalogs? Can expect a messy world to be captured in a clean set of meaningful categories? Or is all information fundamentally uncertain and only partly known? How can we fruitfully combine information retrieval and semantic web approaches? These and other related questions will be discussed at this open format workshop -- the aim is to provide paths for further research to change the way we understand information access today! The workshop will consist of three main parts: Three keynotes to help us frame the problem, and create a common understanding of the challenges: Kevyn Collins-Thompson (University of Michigan); Marti A. Hearst (University of California, Berkeley); and Dan Roth (University of Illinois at Urbana-Champaign). A boaster and poster session with 14 papers selected by the program committee from 21 submissions (a 67% acceptance rate). Each paper was reviewed by at least two members of the program committee. Breakout groups on different aspects of exploiting semantic annotations, with reports being discussed in the final session.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.014 | 0.019 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.003 | 0.002 |
| Bibliometrics | 0.004 | 0.005 |
| Science and technology studies | 0.002 | 0.003 |
| Scholarly communication | 0.012 | 0.018 |
| Open science | 0.004 | 0.005 |
| Research integrity | 0.004 | 0.006 |
| Insufficient payload (model declined to judge) | 0.053 | 0.029 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".