BY ROBERTO PIERACCINI

First let me wish you all a Happy 2008! As usual, this time of the year we thank those members who have accomplished their three year term and we welcome the new members coming on board. Please join me in a more than deserved thank you to Jean-Claude Junqua, Peter Kabal, Hisashi Kawai, Tomoko Matsui, Roger Moore, Shri Narayanan, and Yunxin Zhao for their help and dedication to SLTC. Also very special thanks go to Mazin Gilbert, former chair of SLTC, part of the committee during 2007 as guest member and past chair. Mazin greatly helped me start off and hit the ground running in leading SLTC at the beginning, and always provided wise and useful advice and suggestions for the rest of the year. Thanks Mazin! You'll be missed.

Jont B. Allen, Frederic Bechet, Jean-Francois Bonastre, Timothy J. Hazen, Alexandros Potamianos, Frank K. Soong, and Yannis Stylianou are the new elected members who joined the SLTC this year. Our warmest welcome goes to them, and our thanks for having started to help, even before the beginning of their term, with the ICASSP review process.

Given the large number of papers, the ICASSP review process is quite complex and requires the orchestrated work of the whole SLTC. Like in the past years, SLTC's responsibility concerns the Speech (SPE) and Spoken Language (SLP) processing areas. This year we received 517 papers in SPE and 101 in SLP. The process started in early October with the identification of a rich pool of qualified external reviewers. Until a few years ago the review of ICASSP papers was done only by the SLTC members. Today, with the large number of submissions we receive in SPE and SLP, this cannot be done anymore. Given that we need at least 3 reviews per paper, with a committee of 46 members (including the new ones), this year each member would have had to review an average of more than 40 papers! And in practice it would have been more than that for some members because of the need to match the topic of each paper with the reviewer's expertise. So as we did during the past few years, we had to resort to a pool of external reviewers, 372 this year. Each external reviewer was assigned a number of papers--the maximum number was 19. Each paper was assigned to 3 external reviewers and an additional SLTC member. 4 area chairs (Ciprian Chelba, Brian Mak, Murat Saraclar, and Gokhan Tur) managed the process of making sure all the papers were assigned to external reviewers in their areas of expertise with no potential conflicts of interest, and each paper had at least one SLTC member assigned to it. During the reviewer assignment period, they resolved unpredicted conflicts of interest, reassigned papers to different topics when they were not considered to be in the right area or the right TC, answered questions from the reviewers, and gently reminded reviewers to finish their work on time as the review deadline approached. By December 5th, 2,400 of the expected 2472 reviews were in place, and the area chairs could start the paper acceptance process.

The number of papers accepted depends on the number of sessions that the ICASSP organizers allocated for each area. This year the SPE and SLP got 34 sessions, including lectures and posters, and that determined the acceptance rate. The overall score of papers, averaged over the 4 reviews, is the criterion used to determine whether a paper is accepted or rejected. In case where there is a high variance across the reviewers, the area chairs, and the SLTC members assigned to those papers, try to reconcile the scores by assessing, as best as they can, the quality of the paper and consolidating the results of the reviews. In a few occasions additional reviews were made by the area chairs or by selected SLTC members.

Given the large number of submissions and their overall high quality, it always happens that good papers cannot make the cut. I want to stress the fact that when a paper is rejected, that does not necessarily mean that the quality of the paper is not high. Unfortunately, having to fit a large number of high quality papers in a limited number of slots produces inevitable "casualties" that, often, may appear unfair. We eventually accepted 300 papers (247 in the speech area and 53 in language processing) with an overall acceptance rate of 48.5%, which is in line with the acceptance rate recommended by the ICASSP organization.

But the work of the SLTC committee and of the area chairs does not stop here. The accepted papers need to be grouped in sessions with consistent topics; titles have to be created for each session; and session chairs have to be assigned to them. Then, some papers have to be selected for various awards, such as the "best student paper award", and the "spoken language processing student grant".

All in all the entire process went smoothly and we have learned a few things that will certainly improve it in view of the next ICASSP edition. Speaking of the whole process, I should not forget to thank Lance Cotton of Conference Management Services, the "man behind the scenes" who every year manages the review web site for all TCs and the whole ICASSP review. Lance is always ready to help, to provide new views and data, and to create last minute scripts trying to accommodate the many - sometimes conflicting-requests from the area chairs and members of the SLTC and - I guess-from all the other TCs involved in ICASSP. Thanks Lance!

Although the ICASSP review constitutes most of the work, SLTC engages through the year in a myriad of other activities. Providing nominations for a number of IEEE awards, and making sure that the areas of speech and language are well represented,  is one of them. This year, thanks to SLTC's nominations, Yuqing Gao and Jeff Kuo won the IEEE best paper award, and Prof. Renato de Mori was selected as an IEEE Distinguished Lecturer. Other activities include coordinating the efforts with other organizations in the speech and language technology areas. This year, for instance, we worked with the HLT-NAACL organization for strengthening the representation of speech technology at the annual HLT conference. To that end we invited Mary Harper (from the HLT board) as a guest member in SLTC. Consequently Ciprian Chelba, from SLTC, joined the HLT board.

I want also to mention that, in addition to the traditional face to face meeting of the SLTC in occasion of the ICASSP conference (that was held in Honolulu in April 2007), last year we had another meeting in August in occasion of the Interspeech conference in Antwerp, Belgium. The attendance to both meetings was extraordinary, and we did accomplish a lot.

I cannot cite all the other activities in which SLTC got involved last year because the long list of them is beyond the scope of this welcome message, but I want to say that as preoccupied as I was with accepting the role of Chair last year, I have learned a great deal, I have had the opportunity of working with the great people of the committee who provided an immense amount of help, and yes ... I had a lot of fun.

See you at ICASSP 08 in Vegas!