On Sept. 7-9, 2002, the Center for Gravitational Wave Physics (CGWP) at Penn State hosted a small informal meeting to discuss the results of the Lazarus project and their relationship to the detection of gravitational waves from merging black holes. This meeting, organized by Manuela Campanelli, Sam Finn (CGWP director) and Pablo Laguna, was designed to maximize the rate of information exchange by limiting participation to just a few people, most of whom were already active in the Lazarus and Kudu projects. In attendance were Abhay Ashtekar, John Baker, Bernd Bruegmann, Jordan Camp, Manuela Campanelli, Joan Centrella, Jolien Creighton, Sam Finn, Pablo Laguna, Ben Owen, Deirdre Shoemaker and I.
For those not familiar with the Lazarus project, it is an attempt to model black hole mergers using the limited available numerical evolution only when absolutely necessary. The goal of the project, led by John Baker, Manuela Campanelli and Carlos Lousto, is to seamlessly sew a complete 3+1D, nonlinear numerical evolution in between a suitable early time approximation of the binary system, such as the PN approximation, and a suitable late time approximation, such as the close limit approximation, and thus produce merger waveforms which are suitable for data analysis. At present, the waveforms produced by Lazarus are not of sufficient accuracy to use for the standard data analysis algorithm of matched filtering. This has led to a second project, the Kudu project, whose goal is to define a framework for going from imperfect waveforms to data analysis algorithms, and to implement the framework on the Lazarus waveforms.
As with many ``working meetings'' that are being hosted by the CGWP, the agenda featured brief talks followed by long discussion periods, a format which I like very much. The first day of the meeting focused on the machinery of the Lazarus approach. Manuela Campanelli started with an overview of the Lazarus and Kudu projects. Almost immediately, there was a flurry of questions and discussion which set the tone for the day - a detailed, highly interactive dissection of the Lazarus project. Carlos Lousto's presentation of Lazarus methodology and John Baker's talk on Lazarus results were direction markers which primarily served to guide the general course of the conversation in which we were already engaged. It would be futile to try to detail the flow of this freeform discussion. Let me instead summarize with what I think was the main conclusion of the day - the Lazarus project has developed a framework which must overcome many technical hurdles, and at each of these hurdles, there is room for error to enter into the calculation. At present, much of the knowledge about these errors has been gleaned by the Lazarus researchers project seeing how things go awry when a mistake is made or they adjust a parameter. However, in order to use these waveforms, data analysts will need quantitative error estimates. Providing these estimates will be a long and arduous process, but the Lazarus project members who were present seemed to agree that it was possible and worthwhile.
The second day of the meeting focused on the Kudu project. This day was much more speculative - the Kudu project had barely started when the meeting was held. As a result, the nature of the meeting changed from critical review to exploring new horizons. I gave the first talk of the day, in which I described how one could devise an optimal (in a certain sense) search algorithm for signals about which only incomplete information was available. Much interesting discussion followed on how one could use this to help refine source modeling efforts by concentrating the effort on those aspects of the model whose refinement would lead to the greatest increase in detection rate. Jolien Creighton then outlined the methods that are currently being used within LIGO to search for unmodeled signals, and how these methods might be improved to look for the Lazarus waveforms.
On the final morning, we had a discussion of where we stood and future directions. However, as with all such discussions, it is in the doing that progress will be made, rather than in the discussion. And this leads me to what I believe was the most fruitful aspect of this meeting. Over coffee, dinners, and late into the evenings, a subgroup of us discussed specific ideas that we wanted to explore, and agreed to have weekly teleconferences to discuss the results we have obtained. This group continues to meet (biweekly now) and to discuss results obtained and brainstorm new directions. It is unlikely that this nucleus could have self-assembled without the environment of a working meeting to stimulate it, and it is exactly by providing such opportunities that I feel that the CGWP will have its greatest impact on our community.