How papers were submitted and presented
Big Learning used extended abstracts to bring research into workshop discussion. The guidelines for the 2011 and 2013 editions suggested papers shorter than four pages, excluding references, in the NIPS LaTeX style. Additional details could appear in appendices or supplementary material. Accepted work was presented through contributed talks or posters, connecting the written account to discussion during the meeting.
Those rules belonged to past workshops. The calendar gives their submission and final-version dates, while the 2013 call explains the research scope. The format makes most sense beside the workshop's purpose: the NeurIPS description of the 2013 edition emphasised exchange between database research and machine learning, including implementation studies and the concepts behind system designs.
A compact written account
An extended abstract occupies a middle ground between a brief description and a full-length research paper. In Big Learning, the suggested page length encouraged a concise account of the problem and the work presented. References were outside that suggested length, while appendices and supplementary material provided room for further detail. The 2013 call itself stated a limit of four pages, excluding references; the guidelines expressed the preferred length as under four.
The distinction between those wordings is useful. A call defines the scope and format of a contribution, while guidelines can also express preferences about how to present it. Neither wording meant that every technical detail had to fit into the main pages. The permitted additional material gave projects with lengthy descriptions another place for that detail, keeping the principal account short enough for a workshop's review and discussion.
Identity and earlier publication
The guidelines did not require double-blind review. They asked for author identities to appear in the extended abstract. This described the review process used for those editions; it did not change the expectation that the work would be evaluated. The organisers also allowed topics recently published or presented elsewhere, provided that the extended abstract explicitly declared the earlier presentation or publication.
Work from conferences outside machine learning was particularly encouraged. That permission fitted a series interested in several research communities, including algorithms, databases, distributed systems and scientific applications. A contribution could bring an established result into a different discussion, rather than needing to be wholly new to every audience. Declaring prior publication made the relationship between the workshop contribution and the earlier work explicit.
From accepted paper to presentation
Accepted submissions were assigned contributed-talk or poster presentations. The 2011 papers, 2012 papers and 2013 papers show how those categories appeared in the programmes. A presentation category identified how the work entered the day; it did not by itself describe the paper's technical content. Paper titles and author lists identify the contributions across both formats.
The 2013 guidelines recommended posters smaller than 170 cm × 100 cm. That practical rule concerned the physical presentation, whereas the extended-abstract rule concerned the written account. The two formats complemented each other: the written contribution supplied a compact description, and a talk or poster made the work available for discussion during the event. The 2013 programme placed poster periods between invited and contributed talks.
Workshop papers and conference papers
Wikipedia's discussion of academic conferences describes meetings where researchers present and discuss scholarly work. It notes that written contributions, presentations, keynote lectures and poster sessions can all be part of a programme. The labels “conference” and “workshop” do not establish a universal difference in publication practice. The particular venue's rules matter more than an assumption based on the label alone.
Big Learning's rules were specific: short extended abstracts, declared prior work, and contributed talks or posters for accepted submissions. Its invitations to practical studies and work from other research communities also distinguished its purpose from a requirement for one uniform kind of paper. The workshops combined presentation of contributions with discussion of tools, algorithms and systems. Reading the author format beside the call helps explain that combination.
The 2011 programme also included poster spotlights. These short appearances and the longer poster discussions were separate parts of the day, showing that one paper could enter the programme through more than one presentation setting. The written format was only one layer of participation. The order of sessions and the time allocated to each kind of presentation determined how contributions became part of the workshop's shared discussion.
What peer review contributed
Wikipedia's account of peer review describes evaluation by people with relevant expertise. In scholarly publishing, this evaluation informs decisions about a contribution's suitability for a venue. For Big Learning, a programme committee reviewed submissions. The use of review and the visibility of author names were separate facts: double-blind review is one possible arrangement, rather than another name for review itself.
A short workshop account can therefore be read as a contribution selected for a particular discussion, with a defined format and presentation setting. Its acceptance does not replace examination of the research methods or the evidence behind any result. The glossary explains extended abstracts, poster sessions and related technical terms, and the edition pages place the written contributions within the aims and structure of each workshop.