Guidelines for ICDL Organizers

These guidelines should be read by all ICDL organizers and program committee members.

1. Scope and Uniqueness of ICDL

ICDL is a multidisciplinary conference pertaining to all subjects related to the development and learning process of natural and artificial systems including, but not limited to, artificial intelligence, psychology, neuroscience, and robotics.

This conference series has a visionary goal --- to understand autonomous development in human and higher animals simultaneously in biological, functional, and computational terms, and to enable such development in artificial systems. This conference series has an open-mind policy to foster advances toward reaching this challenging goal.

It is important to keep in mind that the ICDL conference is unique, in comparison with other existing conferences that cover intelligence related subjects. Some conferences focus on human development, with little consideration of biology or computation. Others focus on robotics, and do not consider natural, biological systems, much less their development. Some others focus on the biology of brain development, but without much consideration of how brain development leads to emergence of autonomous, intelligent behavior. Additionally, some conferences take a computational approach to human development, but do not consider the biological mechanisms. Finally, others consider the neural mechanisms of intelligence or the evolution of intelligence from a computational point of view, , without considering the development.

The goal of ICDL is to foster research that integrates all of the above approaches. That is, we foster progressive integration of multiple causalities: from human behavior causality, to computational causality, to brain causality, to developmental causality. Work of this integrative nature represents what is unique about the ICDL conferences, and it is also especially challenging to address. However, all work that is related to the general subjects of development and learning in natural and artificial systems falls into the scope of this conference. Because of the special challenges facing integrative work, care should be taken in formulating review committees and acceptance policies to ensure that reviewers sympathetic to the integrative goals of the conference are included, and to ensure that the special challenges confronting work of this type will be taken into account in the review process. As a general principle, the basis of judgment of a submission should be state-of-art for work of its type. The level of rigor that is possible for some types of work may be different than what is possible for work of other, more established, types.

2. Organization of ICDLs

A proposal of ICDL organization should include the names of the General Chairs, the names of Program Chairs, the venue and approximate date of the conference. ICDL Governing Board makes the decision based on their merit. ICDL Governing Board may also directly invite key people to submit a proposal. The principle of organization of ICDLs is the well established principle of checks-and-balances. All conference organizers have the responsibility for maintaining a standard of ethical conduct with respect to paper (and award) review and selection. This includes enforcing conflict of interest rules and maintaining the confidentiality of reviewers, individual reviews, and program committee members.

2.1 Diversity of organizers

The diversity of ICDL organizers is a key mechanism to serve the cross-disciplinary nature of this conference.

(a) General chairs and program chairs should represent both the "natural intelligence side," such as neuroscience and psychology, and the "artificial intelligence side," such as machine intelligence and robotics. The ICDL Governing Board oversees the diversity.

(b) The diversity of program committee members and reviewers should be promoted by the conference organizers.

2.2 Procedures for fostering ground-breaking research

(a) For some consistency of review standards, papers are reviewed by program committee members overseen by area chairs and program chairs, as explained in Section 7.

(b) For strongly divided outcome of the review process for a particular paper, consultation among the general chairs and program chairs should occur to determine how best to address the discrepancy. Concerns expressed by negative reviews will be given careful consideration, but our policy should be to avoid allowing a single negative review, especially if not fully justified by clear arguments, to over-ride enthusiasm by other reviewers.

2.3 ICDL business meeting

(a) ICDL business meeting is jointly held by the ICDL Governing Board, IEEE AMD Technical Committee, and the conference organizing committee, during the ICDL conference. All the members of ICDL Governing Board, IEEE AMD TC members, and expended members of organizing committee are requested to attend the business meeting.

(b) The documents that need to be voted on at the ICDL business meeting need to be made available reasonably before the ICDL business meeting

(c) The meeting follows the Robert's Rules of Order.

For simplicity, and based on the types of methodologies that are typically employed, we often divide the contributing disciplines into two rough categories: "artificial" and "natural". While we realize that many members of our community span this continuum, this rough division is often helpful in maintaining an appropriate balance of the organizational structure.

3. Submissions

3.1 Full paper submissions

Submissions to ICDL are typically a maximum of 6 pages in length and formatted according to the IEEE conference publication templates. (This ensures that papers can be rapidly included in both the proceedings and the IEEE digital archive).

3.2 Abstract submissions

For encouraging latest results, a second type of submission: a one-page abstract is also accepted. The deadline for the submission of abstracts was significantly later than the deadline for full (6-page) papers, and ideally will occur after the accept/reject decisions on full papers are sent. These offer an opportunity for anyone to attend ICDL and talk about their work. These abstracts are presented as posters in a special session (on a different night than the full 6-page submissions). These are typically reviewed only by the Program Chairs (or by a designated qualified individual). The reviews are relatively light, and only seek to eliminate things that are clearly out of scope for the conference. (In 2008, we rejected only 2 of 32 abstracts.)

3.3 Special sessions

Special sessions can often increase the interest and attendance for ICDL by attracting a sub-field or drawing in a number of senior researchers. Typically, special sessions are organized outside of the typical review process and do not have papers included in the proceedings, though there is nothing that prevents the inclusion of invited material. (Invited talks and speakers are not eligible for conference awards, however.) A chair for the special session should submit a proposal to the PC chairs, who have the final decision on both the approval of the session and the time to be allotted to the session.

3.4 Tutorial sessions

Tutorial sessions are encouraged. As this is a cross-disciplinary field, well selected and well run tutorials can increase the number of participants and grow this community. The organizers are encouraged to run tutorials free of charge to registered participants using classroom facility that does not increase the financial burden of the conference. However, all the tutorial participants are required to register and pay registration fee.

4. Conference Organization Team

4.1 General Chairs

The general chairs are responsible for the general organization of the conference, including budget, venue, and overseeing program chairs, publicity chairs, publication chairs, tutorial chairs, etc. General Chairs and ICDL Governing Board Chair jointly appoint the Program Chairs, with input from past Program and General Chairs.

4.2 Program Chairs

Two program chairs should be selected each year, one from the "natural intelligence" side and one from the "artificial intelligence" side.

The program chairs have responsibility to construct all aspects of the scientific program. This includes call-for-papers, web construction, paper submission and review, and program schedule.

The program chairs should keep the General Chairs informed regarding the progress of the scientific program, and get inputs from the General Chairs regarding special sessions, invited talks, award committee and conflict of interest, among others.

4.3 Publicity Chair

The publicity chair is responsible for a wide distribution of information about the conference to a wide array of multidisciplinary fields, including psychology, neuroscience, cognitive science, artificial intelligence, neural networks, machine learning, robotics, and pattern recognition.

4.4 Publication Chair

The publication chair is responsible for interfacing between IEEE that deals with the publication of the conference material and the program chairs, including construction of web, the printing of call-for-papers, the production of proceedings with abstracts, and the production of CD proceedings.

4.5 Program Committee

The primary task of the Program Committee is to ensure that submissions receive a timely, accurate, and fair review. The Program Committee is composed of: (a) Program Chairs; (b) Area Chairs; (c) Program Committee Members. Program Chairs appoint Area Chairs. The number of Area Chairs is approximately balanced in both natural and artificial sides. Area Chairs suggest candidates for Program Committee (PC) Members. Program Chairs appoint PC Members based on Area Chairs' suggestion and on previous years' list of PC Members (Previous irresponsible PC members should be excluded).

5. Review Process

Based on the nature of a submitted paper, Program Chairs assign a primary Area Chair and a secondary Area Chair. The primary Area Chair should assign at least two PC members to each paper, and the secondary Area Chair should assign at least one PC member to each paper. Area Chairs are responsible for ensuring that the PC members submit reviews on time.

PC members should review the assigned papers themselves, and can occasionally solicit help from senior PhD students or collaborators but they are still responsible for the review quality. The primary Area Chair provides a summary review and makes accept/reject recommendation.

Program Chairs make the final decision. In case Area Chair's recommendation is not retained, a discussion must take place among the Area Chair and the Program Chairs, and if necessary, involving the General Chairs.

After a list of accepted papers has been generated, the Program Chairs will then make a decision on how each accepted paper will be presented: either as an oral presentation or as a poster presentation. This decision often reflects a combination of paper quality (as reflected in the reviews), novelty, and the expected size of the audience. For example, papers that are very high quality but will be of interest to only a small segment of the ICDL attendees will typically be presented as a poster. ICDL is a single-track conference.

6. Keynote Speakers

The number of keynote speakers is determined each year by the General Chairs based on available budget. Invitation of keynote speakers is made preferably through a consensus decision among the General Chairs and Program Chairs, and through voting in case of no consensus.

7. Timeline

The application of each conference for financial sponsorship and the budget must be submitted to IEEE CIS at least 2 years before the date of the conference as required by IEEE CIS. The venue contract must be signed timely to guarantee the availability of the rooms. Any uncommitted rooms must be released back to hotel before the hotel deadline, to avoid any room penalty.

This is a sample timeline of the review process, normalized to the day that full paper submissions are due.
Day Deadline
-24 Special session proposals due
0 Full paper submissions due
5 Program Chairs should have papers assigned to Primary and Secondary Area Chairs
10 Area Chairs should have papers assigned to 3 or more PC members
38 Reviews due
40 Review sent to authors for rebuttals
47 Rebuttals due
54 Summaries and recommendations due from Area Chairs
61 Program Chairs decisions made
63 Notification of accept/reject for full paper submissions
70 Poster abstracts due
77 Notification of accept/reject for abstracts
84 Camera-ready papers due
85 Author early registration in order for their papers to be included in the program
90 Final program announced
95 Deadline for advance registration and hotel room blocks (may be adjusted to the hotel contract). Never agree to pay penalty for unoccupied rooms in the hotel contract.
96 Final version of paper proceedings (with all abstracts) and CD proceedings
110 Conference start

Note that these deadlines leave relatively little time for some tasks, and should be taken as a latest-possible deadline. Care should also be taken to schedule both the conference and the submission deadlines so that they do not conflict with other deadlines that typically occur in the spring for related meetings (CogSci, AAAI, and IJCNN are particular conferences that often have close deadlines to ICDL.)

8. ICDL web sites

All the ICDL web sites should be kept permanent. The host institution acquires a domain name like http://www.icdlxx.org/ for the year of the conference. The trial version and the permanent site should be maintained under the host institution's domain name, such as http://www.hostinstitition.edu/icdlxx/

The web site of each ICDL conference should have a link to the ICDL history site, currently at http://www.cogsci.ucsd.edu/~triesch/icdl/

9. Proceedings

The Publication Chair deals with the archiving of papers at IEEE Xplore and the production of the conference program brochure. Accepted full papers can be included in the conference CD proceedings, but 1-page abstracts (which have not had a full peer review) cannot.

10. Awards

Each review form will include a separate entry to indicate whether the work is of award quality. All candidates for award are ranked by the number of award recommendation that each paper has received. A proportion of top-ranked papers are considered by the award committee. All work within the scope of the conference is eligible for awards. As typically conference rewards tend to be given to papers that improve solutions to well understood problems, when other parts are equal between two candidate papers, priority should be given to papers that provide more contribution to the visionary goal and uniqueness of ICDL conference.

The program chairs also have the responsibility for appointing a selection committee for conference awards. The program chairs may or may not be part of this committee at their discretion (and within the bounds of conflict of interest).