Guidelines for Proposals

Selection of Conference Proposals

(Updated in 2019.)

Proposals for the organization of the conference are solicited two years in advance, i.e., the decision on proposals for organizing the conference in year n+2 are made during the conference of year n. In the following, “the conference” always refers to the conference n, i.e., to the conference cycle in which the proposal is made and decided.

The Proposals Chairs have to issue a call for proposals in a timely fashion before the conference (the deadline should be ca. 2 months before the start of the conference).

Candidate organizers are expected to contact the Proposal Chairs as early as possible, and are expected to be in regular contact with the Proposal Chairs during the writing phase of their proposal. During the writing phase, it is highly recommended that the instructions indicated below are taken into account, as well as further recommendations given by the Proposal Chairs. The Proposals Chairs are the prime contact for the proposers, but it is important to note that they do not have the mandate for making any decision.

All proposals will be discussed among the members of the steering committee, and feedback will be given to the proposers. The goal is to ensure that the final proposal contains all the necessary information to enable the community to make a decision.

The Steering Committee has the mandate to reject a proposal if it thinks that it does not meet the quality constraints of this conference series. This pre-selection is made only to ensure a minimal quality of each individual proposal. The steering committee does not have the mandate to make a comparative evaluation among different proposals.

All proposals that are approved by the steering committee will be published on the community Website at the latest one week before the start of the conference. The final decision among competing proposals is made in the community meeting of the conference, in a ballot vote among all people attending the community meeting. In the case of a tie, the vote has to be repeated once among all tied proposals. If there is still a tie, the steering committee will make the final decision.

If there is only one proposal, it also has to be presented in the community meeting, and must be approved by the community. The approval of the community can only be waived if there is no opportunity to get it, e.g., if there is no proposal at the time of the conference that was approved by the steering committee. In this case, the steering committee must actively solicit a proposal and may accept it without the explicit consent of the community.

Preparation of The Proposal

A proposal should provide (at least) the necessary information to allow the community to assess and compare different proposals according to the criteria listed in the following. We also give some recommendations about the implementation of these options, but these are not compulsory.

Chairs

The conference chairs are an important selection criterion. The people that you nominate will not only have to organize the proposed conference, but will also be eligible for the steering committee for three years, and have thus an important impact on the future of this conference series. For this reason, e.g., we generally discourage to nominate people that have not been active in the community in the past.

There are no binding rules for nominating the program chairs, but here are a few criteria that are of concern to the community:

Scientific Merit

Clearly, we expect the chair persons to have a good overview of current research in machine learning and in data mining. This should be evident from their publication history in these fields.

Balance

The program chairs should contain members from both communities, machine learning and data mining. Typically this also implies that the chairs come from more than one group, but this is not required if the group is sufficiently broad. Gender balance is also appreciated by the scientific community. See the list of women in machine learning and data mining.

Community Recognition

A chair person should be part not only of the machine learning or data mining communities at large, but of the ECML PKDD community in particular, so that the community can be reasonably sure that the person is familiar with the traditional lay-out of the conference. The chair persons should have regularly attended the conference in the past, and, ideally, published there before. Previous active involvement (e.g., as a tutorial presenter or workshop organizer) is also a plus.

Number

(Updated in 2019.) The distribution of the tasks to chair members (e.g., general chair, local chair, program chair, journal track chair) is left to the organizers, but the proposal should contain a specification of the structure. The ideal number of chairs responsible for a given task is not larger than 2. It is not necessary that all persons meet all criteria mentioned earlier, but the more the better. However, each criterion should be met by at least one person of the committee. We recommend that one of the General Chairs is also a local chair.

Scientific Background

Discuss how you plan to ensure the scientific quality of the conference. Points that may be considered here are:

  • Do you have any ideas for improving the quality?
  • Is there a local ML or KDD group/community that will be involved in the organization and funding?
  • Do you plan to co-locate with other conferences?

Budget

The budget is a crucial issue on your proposal. The community expects a first rough estimate of the registration fees, which should go beyond a simple statement of intention (such as “We do not plan to increase the registration fees.”). Provide numbers on all your expected expenses. The steering committee can provide numbers on standard costs such as the production of proceedings to help you in that endeavor. Ideally, you show something like a pie chart.

In general, we prefer well-organized events, but also low registration fees. We are a scientific community and typically work in universities. Typically, ECML PKDD conferences take place in universities because they provide an ideal infra-structure at a low price. If you want to deviate from this (e.g., by organizing the event in a conference center or hotel), these additional costs should be justified and their impact on the registration fees should be taken into account. Similarly, if you intend to employ the service of a professional conference organizer.

In order to keep registrations fees low, it is also a good idea to attract sponsors. It is good if you can already name a few (realistic) potential sponsors.

Venue

Please provide detailed information on the venue of the conference. Important parameters are

  • number and sizes of conference rooms
  • walking distance between conference rooms
  • audio-visual equipment
  • seating layout (e.g., visibility may be low in the back of rooms with flat rows)
  • seating equipment (desk, power plug, …)
  • internet connectivity
  • total cost of the venue (cf. Budget)
  • facilities for poster sessions and demos
  • accessibility for handicapped persons

Information about the location of the venue is, of course, also appreciated, but you are strongly encouraged to keep the touristic information short. Typically, it is a plus if the location is easy to reach, has good connectivity, cheap flights, etc.

Accommodation and Infrastructure

Please provide detailed information on number of hotel rooms, expected prices, and distance from the venue. Is there low-cost, quality housing available for attendees (especially graduate students)? How far from the meeting rooms? Where will attendees eat? Please estimate costs for meals and lodging.

Briefly summarize travel information, in particular if the conference location is not directly reachable with an international airport.

Provide information on infra-structural services that you plan to offer for the conference participants (e.g., child daycare).

Schedule

(Updated in 2021.)

It is not necessary to provide a detailed schedule at the time of the proposal, but at least the conference dates should be mentioned. Here are some suggestions and constraints on the schedule (but see also key dates on previous editions):

  • Conference website online: June/July previous year (p.y.)
  • Approaching potential sponsors: beginning of October (p.y.)
  • Approaching potential invited speakers: October (p.y.)
  • Journal track submission deadlines: cutoffs in October (p.y.), December (p.y.), February
  • Conference tracks submission deadline: end of March / beginning of April
  • Demo track submission deadline: end of April / beginning of May
  • Industry track / PhD forum submission deadlines: mid June
  • Conference dates: mid September
  • First call for papers out: at least three months before the submission deadline
  • First call for proposals out: at least five months before the submission deadline
  • Workshop/Tutorial/Discovery Challenge proposal deadline: beginning of February
  • Workshop/Tutorial/Discovery Challenge acceptance notification: beginning of March
  • Workshop/Discovery Challenge CFP out: three months before the submission deadline
  • Workshop/Discovery Challenge submission deadlines: mid June
  • Early bird registration deadline: after the last notification deadline.
  • Camera ready deadlines for proceedings (pre-proceedings): at least nine weeks before the conference (final material should be sent to Springer at least six weeks before the conference)

Conference Format

ECML PKDD has been a dynamic conference that has undergone quite a few changes in recent years. Nevertheless, some key parameters (e.g., the conference schedule) have remained stable over the years. The general structure, typically very much appreciated by the community, consists of:

  • conference sessions from Tuesday till Thursday
  • workshops and tutorials on Monday and Friday, free for conference participants
  • discovery challenge
  • invited speakers
  • poster sessions (for all accepted papers) with buffet
  • plenary sessions
  • coffee breaks
  • conference banquet
  • child daycare
  • social events (e.g. guided tours)

Publication

(Updated in 2019.)

  • the organizers will be responsible for maintaining the journal track, with publications in the DMKD and MLj journals.
  • long papers (16 pages in LNCS-format) are strongly preferred
  • printed workshop proceedings and tutorial notes are not necessary
  • short papers are discouraged
  • videos are generally liked and frequently used

Reviewing Process

  • author feedback is appreciated
  • conditional acceptances are less appreciated
  • no double submissions with other conferences should be allowed

Conference Program

  • poster sessions for all papers are generally liked
  • short walking distances between conference venue and hotels are appreciated
  • parallel sessions in different buildings should be avoided if possible
  • avoid to put award-winning papers in simultaneous parallel sessions
  • long bus trips are not a good idea

General Comments

As the conferences have now grown together, we must be careful not to lose the balance between both communities. However, our community changes, and so might our preferences. Creative ideas for improving the conference organization have always been welcomed by the community, and could add to your proposal. We also know about the importance of exploitation and exploration. Feel free to try out new things, but plans for major changes should be first discussed with the steering committee.

The Presentation

All proposals that have been approved by the steering committee will be available at the community Website one week before the conference and have to be presented at the conference’s community meeting.

The time for the presentation is strictly limited to 10 minutes (excluding questions) and should provide all of the above details, so that the community has all the information for choosing the better proposal based on merits. The time is limited, so touristic information should be kept short.