Vol. 1, #2 | | November 1990 /| The /| / | / | ---------/ | /---- ISSNNet ----/ | /------------ |/ |/ % % %%%%%% % % %%%% % %%%%%% %%%%% %%%%% %%%%%% %%%%% %% % % % % % % % % % % % % % % % %%%%% % % %%%% % %%%%% % % %%%%% % % % % % % % %% % % % % % % % %%%%% % %% % %% %% % % % % % % % % % % % %%%%%% % % %%%% %%%%%% %%%%%% % % %%%%%% % % ----------------------------------------------------------------------------- A quarterly publication of the International Student Society for Neural Nets All correspondence should be sent to one of the following addresses: Surface: ISSNNet, P.O. Box 557, New Town Br., Boston, MA 02215, USA Electronic: issnnet@park.bu.edu (ARPANET) Those of you who wish to contribute to future issues of this newsletter should send e-mail to , or to the address for the specific column for which you wish to write. All inquiries and comments related to the Newsletter in general should also be addressed to . All other inquiries should go to the main addresses listed above. We hope that future issues of the newsletter will contain many of your contributions, to reflect the ISSNNet ideal of promoting cooperation and an open exchange of information. ----------------------------------------------------------------------------- || ISSNNet Feature Article -- Submissions to: || ----------------------------------------------------------------------------- Hiring New College Graduates: One Employer's Perspective by Patrick K. Simpson** Embarking upon a career that will apply all the wonderful and important things that have been learned during your education is both exciting and frightening. The excitement comes from finally getting paid to do what you have so diligently worked toward for several years. The fright comes from what you do not know. You are probably asking yourself, What is an employer looking for in a candidate employee? How can I optimally present myself so that I am seen for the valuable person that I am? Below are some suggestions and hints about things that I look for in a candidate employee. I have broken the discussion into two parts; the resume and the interview. Please keep in mind, this is only one man's opinion. THE RESUME ---------- A person's resume is their life-story on a single sheet of paper. When I look at a resume, I typically scan it quickly - maybe 30 seconds. If I see things of interest, I will study it further. For this reason, it is very important to have key information readily accessible. There are several things that I look for when I review a candidates resume. (1) DEMONSTRATION OF TECHNICAL COMPETENCE The first thing I look for is the technical qualifications of the candidate. This should come as no surprise. If the candidate has a technical background in the same (or similar) area that the position requires, then other factors may be considered. (2) WRITTEN COMMUNICATION SKILLS When considering a candidate, the single most important aspect is the ability for that person to demonstrate effective communication skills. Although the ability to orally communicate can not be fully evaluated in the resume (this must be done in the interview), a candidate's ability to communicate in writing can. There are several indicators that I use to determine good written communication skills. The first is the resume itself. If the candidate has succinctly presented a one-page description of their capabilities and qualifications in a well organized and grammatically correct manner, then this tells me this person has potential. Be aware: misspelled words can hurt and typos can be deadly. Second, I look for papers. The more advanced the degree, the more papers I would expect to see. If the candidate has a Bachelors degree, I would expect possibly one or two project papers that were associated with course-work. If the candidate has a Masters degree, I would expect to see a conference publication and a thesis. Finally, if the candidate has a doctorate, I would expect to see several conference publications, at least one refereed journal article, and the doctoral thesis. I usually ask the candidates to send me a few of their best papers prior to the interview. Doing this without asking is a very positive indicator as it shows pride in authorship as well as foresight. (3) PRIOR WORK EXPERIENCE It has been said that there is no better teacher than experience (at least that is what my mother always told me). This is especially true when looking for a job. One of the best ways to land a job when you get out of school is to have a job before you leave. I look for student interships, part-time work, and research positions on the resume. I don't necessarily place off-campus work ahead of on-campus employment, but I do look for work that is similar in nature to that I have a need for. A candidate with signal processing experience would probably be better suited for a company dealing with radar applications than one dealing with human factors. Attaching letters of recommendation and/or references from previous and/or current employers along with your cover letter is always good. (4) KNOWLEDGE ABOUT THE COMPANY Very few candidates take the time to research a company before they send their resume. This can be usually done with a single call to the human resources department of the company and it carries significant impact. If a candidate collects information about a company, then it is possible to add a subtle slant to the resume to fit the company's focus. For example, if you are applying to a company and you discover that the group you are interested in working for has been working on a specific type of radar system, you can phrase your talents with respect to the capabilities necessary to work with such systems. This information is also useful in the cover letter and the interview. In addition, if a candidate can call the person that will receive the resume and briefly introduce themselves, then that person has improved their chances from the start. (5) UNIVERSITY ATTENDED AND DEGREE EARNED Note that I have listed this fifth. This is not accidental. I am looking for discriminators when I am looking at a resume. Why is this person better than the other five resumes that I have seen this week. In my opinion, communication skills, prior work experience, and company knowledge are those types of valuable discriminators that carry much weight. This does not diminish the importance of the degrees. THE INTERVIEW ------------- The interview is the candidate's opportunity to convince the employer that they are the right person for the position. There are several elements that I look for when conducting an interview. (1) ORAL COMMUNICATION SKILLS One of the most important elements in the interview is the demonstration of good oral communication skills. I look to see if the candidate listens, answers succinctly, and uses proper language. If a person gets easily flustered it is not a serious problem--everyone gets nervous. If the candidate tends to ramble or of the candidate seems to easily stray from the subject, these are often seen as negative attributes. I usually pick one technical element from the resume and ask the candidate to provide a brief explanation of that work. Clear and understandable explanations have a positive impact. (2) DRESS TO IMPRESS First impressions are important. Dressing professionally is a key to a successful interview. I am always reminded of a pair of pictures that I saw in a magazine article many years ago. In the first picture there was a man in a long untrimmed beard, long hair, a dirty tee-shirt, torn blue jeans, and sneakers. In the second picture was a man with short neat hair, no beard, a three piece suit, and loafers. Under both pictures was the caption "Which one would you hire?". The punch line was that these pictures were of the same person. It might not be fair that people assess you by appearance, but the reality is that it happens. (3) ASK QUESTIONS I am always impressed by a candidate that asks questions and shows a genuine interest in the position available. Questions demonstrate that the candidate has been listening and has been thinking. Good questions to ask are, What would my day-to-day working activities be? What are the goals of the program I would be working on? Who would be my direct supervisor? etc. Please remember that I might be looking for things that are completely different than another perspective employer. I hope that these comments are helpful to those students that are, or will be looking for a job in the near future. _____________________________________________________________________________ ** Pat Simpson is currently employed at General Dynamics, Electronic Division, PO Box 85310; MZ 7202-K, San Diego, California 92136, USA. The views expressed here are solely those of the author and they do not necessarily reflect those of General Dynamics Corporation or any of its subsidiaries. General Dynamics is an equal opportunity employer. ----------------------------------------------------------------------------- ||-- ISSNNet Status Report --|| ----------------------------------------------------------------------------- by Paolo Gaudiano The last issue of the ISSNNet newsletter featured a personal perspective on the history of the field of neural networks. The purpose of that article was to show "where we are coming from," and to suggest how a student society such as ISSNNet could be useful to the field. Since that time the society has had its debut in the "real world," and the results have been extremely positive. This note summarizes the projects we have worked on this summer, and shows how the experience we gained has helped cast the purpose of ISSNNet into more concrete terms. But first, we'd like to apologize for the slow response to those of you that have already paid the $5 membership fee and have gotten no response. We have not forgotten you, and you should get a membership card and list of all members very soon. The first newsletter was followed by an effort to organize ourselves for the conferences in San Diego (IJCNN) and Paris (INNC). The organizers for both of these conferences had donated a booth for ISSNNet materials in the exhibit hall. The main purpose of the booth was to sell T-shirts, sign new members, and gather information on existing academic programs and job opportunities. The society had an informal meeting/dinner in San Diego that attracted about 60 people. We signed on almost 100 new members, and made ourselves known to many people in the neural network community. Similarly, the Paris conference attracted almost 100 more members, and established further contacts with a number of European academic and professional institutions. We took advantage of these conferences to publicize two of our projects: To compile a listing of academic programs around the world that are involved with neural networks, and to provide a profile of neural network employment opportunities for recent graduates. Specifically, we asked people to fill out one of our questionnaires, and the results are summarized later in this newsletter. In addition, we had collected a number of student resumes and submitted them to companies that have hired or are looking to hire recent graduates with neural network experience. In addition to these projects, we broadened our network of ISSNNet "governors" throughout the world. The purpose of each governor is to act as liaison between ISSNNet and local members, as well as to encourage formation of local groups for various activities. This should promote an exchange of information among individuals in different fields and/or at different geographical locations. We are currently working on a complete membership list that will be mailed to official members, and will also be broken down into geographical areas by the governors. These examples are helping us to define in concrete terms what it is that the society can do, and how it fits in the scheme of neural network research. As many people have noticed, the "field" of neural networks is a highly fragmented one. People in engineering are doing different work from those in psychology, or physiology, or computer science. Some of the work is purely theoretical, some is purely applied, and a lot is somewhere in between these extremes. Many companies have little ideas on where to find good candidates for neural network positions, and the hiring criterion is often based on programming experience. This problem seems to stem from two distinct phenomena: First, there are extremely few academic programs devoted specifically to neural networks, which makes it difficult for companies to assess an individual's qualifications based on where he or she received a degree. Second, there is no real "standard" in neural networks. Many applications make use of back propagation or other gradient descent models, but as of yet there exist no standard benchmarks or other methods to decide which model is best suited for a particular problem. As a result, companies often hire people simply to investigate what neural network technology is available, and how it could be applied to their products or research. The student's perspective on these problems was one of the reasons for the creation of ISSNNet. Most students are only exposed to neural networks through a single established field, which makes it difficult to approach neural network models that are formulated in the jargon of a different field. As a result, students may become highly proficient in a rather narrowly-defined area, but may have difficulty finding employment opportunities that take advantage of their learned skills. It is important to stress that ISSNNet is not an organized structure with well-defined purposes and the means to carry them out. Instead, ISSNNet should be regarded as an emergent structure based on the desire of individuals to improve the current state of affairs. The society started when a handful of students realized it was important to establish a common point of contact among students and between students and other organizations. All tasks we have undertaken have been motivated by people's ideas about programs that could be useful to the student community: We are trying to establish a new link for information exchange through various communication media by using computer mail when we can, and surface mail otherwise. We want to provide a centralized source of information for prospective students, so that they may be able to better select the academic program that best suits them. We are establishing ourselves as an important liaison between would-be-employees and employers, by providing the companies with information on recent graduates, and vice versa. One of the original reasons for the creation of the society was to do something about the declining level of assistance for students that want to attend neural net conferences. In the years since the first ICNN and INNS conferences, the average student registration fee has been increasing, while the students get less and less for their money. One of the reasons for this trend is the unwillingness to give students a good deal (and thus losing money) when it is unsure how many attendants the conference will draw. As part of our commitment to this issue we may oversee the organization of volunteers for some future IJCNN conferences, and we have been involved with negotiations to insure that conference volunteers get adequate reward for their work. We are also hoping to organize an entire session dedicated to student presentations at some of the future conferences. From a more practical standpoint, we have officially become a non-profit corporation in the state of Massachusetts, and we are planning to apply for tax-exempt status with the IRS, which makes student support more attractive to companies because donations are tax-deductible. At the same time, we are looking into the possibility of establishing international chapters to better handle all the members. We plan on using whatever profits we make to support students in various ways, including direct monetary contributions. As the society grows in many different ways, the tasks at hand are quickly outgrowing our ability to keep up with them. We are finding out, for example, just how much detail is involved in setting up a tax-exempt, non-profit corporation, and to keep track of all the memberships and the programs. One of our short-term goals is to have official elections before the end of the year. We have drafted a set of bylaws that describe in great detail how the society will be run, and how it will self-organize for future years (send e-mail to for an electronic copy, available in plain text, PostScript, or LaTeX). WE NEED YOUR HELP!!! As we have mentioned, ISSNNet is based entirely on individuals' desire to organize projects that help the Neural Net student community. Continued success depends on the involvement of a lot of people. In this newsletter, you will find the description of a number of positions for which we need volunteers. If you would like to be involved, let us know. Among other things, this is a great opportunity to get in touch with other students and professionals that are interested in the development of the field of neural networks and the role students play within it. ----------------------------------------------------------------------------- ||--|| EMPLOYMENT FORUM - Submissions: ||--|| ----------------------------------------------------------------------------- by Karen Haines At both conferences this last summer (IJCNN in San Diego and INNC in Paris), employment questionnaires were passed out to many companies attending the conference. The first part of this column is a summary of these questionnaires. Most of the companies are from the U.S.A., but we hope that as ISSNNet grows in the rest of the world, we will receive more information from other companies. In regards to US companies hiring foreign students, you must contact the company. In general, it will be your responsibility to get a Visa or work permit. If you contact one of these companies, please be sure to mention that you heard about them through ISSNNet. If you have knowledge of any companies hiring please send electronic or surface mail, and we will be glad to list them in the following newsletter. Good luck to those seeking employment. ............................................................................. NOTE: In some cases the person who filled out the form may not be in charge of personnel, but may refer you to the correct person to contact within that company. COMPANY: AEA Technology ADDRESS: Harwell Laboratory, Oxfordshire, OX11 ORA, United Kingdom POINT OF CONTACT: Personnel Department, or RCF McLatchie COMPANY'S EFFORTS IN NEURAL NETWORKS: Development of Netvision tool for data analysis. Participation in ESPRIT and other joint projects. Robotics and other projects. HIRING THIS YEAR? Yes HIRE IN 1991? n/a TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: High quality graduates interested in research type applications may be considered. COMPANY: Allen-Bradley Co. ADDRESS: Dept 756A, 1201 S. 2nd St., Milwaukee, WI 53204, USA POINT OF CONTAC: Dr. G.W. Davis COMPANY'S EFFORTS IN NEURAL NETWORKS: Product motivated R&D. HIRING THIS YEAR? Yes HIRING IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: This company has hired neograduates in the last two years. Personnel hired had experience and/or knowledge of Object Oriented Programming and Advanced Control Theory (ANN). COMPANY: AMOCO Production ADDRESS: PO Box 3385, Tulsa OK 74102, USA POINT OF CONTACT: Donald H. Mitchell COMPANY'S EFFORTS IN NEURAL NETWORKS: Fielded seismic event detection application. Working on other seismic, geological log, and gas chromatography problems. HIRING THIS YEAR? Yes HIRE IN 1991?Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Research positions are open within the AI group for persons with a Neural Net/AI related MS or Ph.D degree and/or a demonstrated research ability. Researchers currently in the group are investigating automated planning for complex data analysis, neural networks for seismic processing, human-computer interfaces for planning systems, and object oriented database support for knowledge engineering. COMPANY: FBI - Engineering Research Facility ADDRESS: Building 27958A, Quantico, VA 22312, USA POINT OF CONTAC: John Hoyt COMPANY'S EFFORTS IN NEURAL NETWORKS: Neural Network applications in signal processing. HIRING THIS YEAR? No HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? No COMMENTS: Looking for EE's and ME's. COMPANY: For a New Social Science ADDRESS: 10 Dent Drive, Radford, VA 24141, USA POINT OF CONTACT: Sam Leven - PHONE: (703) 633-0665 COMPANY'S EFFORTS IN NEURAL NETWORKS: Research in theoretical networks and theories of Bargaining and Mind. Also looking into Distributed Intelligent Environments. HIRING THIS YEAR? No HIRE IN 1991? n/a TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Company may have a position in dynamical system research. Programming experience and neuropsychology a plus. COMPANY: Ford Areospace ADDRESS: 1322 Space Park Dr. POINT OF CONTAC: Richard Moe COMPANY'S EFFORTS IN NEURAL NETWORKS: Presently involved in turn-key neural system. HIRING THIS YEAR? Yes HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Possibly COMMENTS: Ford Aerospace is currently in the marketing phase of the system they developed. Acceptance of the system will reflect the number of employees that will be hired. COMPANY: HNC ADDRESS: 5501 Oberlin Dr., San Diego, CA 92121 POINT OF CONTACT: Laura Jones COMPANY'S EFFORTS IN NEURAL NETWORKS: R&D, applications, software tools, and hardware. HIRING THIS YEAR? Yes HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Positions in Software Engineering for development of neural network applications and software packages. Most positions require Masters or PhD, and neurocomputing experience as students. C programming skills required; Experience in neural networks, signal processing, or applications area a plus. COMPANY: K.I.A. ADDRESS: Grundtvigsvejg, 8723 Losning, Denmark. POINT OF CONTACT: P. Lyngroes COMPANY'S EFFORTS IN NEURAL NETWORKS: Real applications for production industry. HIRING THIS YEAR? Yes HIRE IN 1991? n/a TEMP. SUMMER EMPLOYMENT? n/a COMPANY: KTAADN, INC ADDRESS: 1340 Centre St #202, Newton Centre, MA 02159, USA POINT OF CONTAC: Donald Frankel COMPANY'S EFFORTS IN NEURAL NETWORKS: Problem solving: Software simulation of Backpropagation on Macintosh, Nestor Learning system on PC. HIRING THIS YEAR? No HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Bachelor or Master's level student adept with Macintosh Toolbox and interface, neural network expertise a plus. COMPANY: Lockheed ADDRESS: 0/91-10 B/282, 3251 Hanover Street, Palo Alto, CA 94304, USA POINT OF CONTACT: Dr. Donald Specht COMPANY'S EFFORTS IN NEURAL NETWORKS: Basic Researchand Applications HIRING THIS YEAR? No HIRE IN 1991? Maybe TEMP. SUMMER EMPLOYMENT? n/a COMPANY: Naval Ocean Systems Center ADDRESS: Research Branch, Code 512, PO Box 997, Kailua, Hawaii 96734, USA POINT OF CONTAC: Patrick W. Moore COMPANY'S EFFORTS IN NEURAL NETWORKS: Biological sonar modeling, sonar signal processing. HIRING THIS YEAR? No HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Possible postdoc available in Neural Networks. COMPANY: NeuralWare, Inc. ADDRESS: Penn Center West, Building IV, Suite 227, Pittsburgh, PA 15276, USA POINT OF CONTACT: R. Jane Klimasauskas COMPANY'S EFFORTS IN NEURAL NETWORKS: Design neural computing development systems. Provide neural net training courses. Develop applications using neural nets. HIRING THIS YEAR? yes HIRE IN 1991? n/a TEMP. SUMMER EMPLOYMENT? No COMMENTS: Seeking software engineers and Neural Net research engineers for application development. COMPANY: Phillips Electric ADDRESS: 4126 St. Clair Ave., Cleveland, Ohio 44103, USA POINT OF CONTAC: Lorry Wagner COMPANY'S EFFORTS IN NEURAL NETWORKS: Time Series Analysis/ Forecasting HIRING THIS YEAR? No HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Two positions, MS/Ph.D, minimum of 1 year experience in neural networks. Additional experience in signal processing, programming, or statistics. COMPANY: SAIC - (Science Applications Intelligence Corp.) ADDRESS: 10760 Campus Point Dr. - M/S 63, San Diego, CA 92131, USA POINT OF CONTACT: Mary Louise Dunford COMPANY'S EFFORTS IN NEURAL NETWORKS: Research & Development focusing on High Tech solutions to problems using neural networks. Company is currently researching uses of neural networks in programs and products. HIRING THIS YEAR? Yes HIRE IN 1991? Yes TEMP. SUMMER EMPLOYMENT? Yes COMMENTS: Company has many openings for various positions is neural networks. Minimum requirement is BS in Computer Science, Engineering, or related area of studies. Products are hardware and software oriented. Request for Proposals: ---------------------- We were recently contacted by Simon Cumming of British Airway's Artificial Intelligence Unit, concerning a small number of sponsorships that will soon be available. Mr. Cumming is presently conducting a study to determine the best use of available funds in the area of forecasting. In particular, Mr. Cumming is investigating the use of neural network techniques to develop applications concerning forecasting or data analysis and modelling in a dynamic commercial environment. Though British Airways is based in the United Kingdom, it may be possible in some cases for the work to be done in other places, providing a knowledge engineer from British Airways would be able to visit for review of the project. Resumes may be sent to: Simon Cumming, AI Unit (s590), Unit 2, Hatton Cross Centre,PO Box 10, Hounslow, TW62JA, England. We suggest that in addition to sending your resume, you include a cover letter suggesting ways neural network techniques can be applied to forecasting. RESUME SERVICE UPDATE --------------------- In general the concept of a resume booklet was eagerly accepted by employers seeking to establish a link to recent graduates with studies in neural networks. In first starting the resume service, the main objective was to have a booklet of resumes of recent or soon to be graduates available to prospective employers at the NNet conferences. Initially a standardized resume format was not considered, however after reviewing the resumes received we realized that a format may be necessary. Instead of passing out the resumes as a product that the society would offer employers, resumes were circulated for review and comments, and of course if any employer wanted to contact a student they were more than welcome. Comments received from employers confirmed the speculation that a standardized format is necessary in order for the resume booklet to be successful. Two comments made repeatedly follow. 1) Lack of a Career Objective or Focus on Desired Employment: Many of the resumes received lacked the focus that employers are looking for in resumes. It may be difficult to write a resume not slated towards a particular position opening at a known company, but maintaining some sort of objective about the position you are seeking is crucial in making your resume stand out among others. Typing up your curriculum vitae and pages of publications without an objective as to what sort of employment you are currently seeking is leaving too much guess work for any employer to deal with. 2) Resumes were Too Long: Although there was A REQUIRED MAXIMUM OF 2 PAGES many of the resumes received were 3 or 4 pages long. Not only are long resumes usually boring, they often provide a lot of information that has nothing to do with getting a technical job. In addition, information is usually repeated. Another reason for a 2 page maximum is that it is expensive to reprint the resumes. Over 200 copies of resumes were made for distribution. A way to cut down the information that is not really necessary on the first pass of the resumes, is to add a line at the bottom of the resume: "Publications, Curriculum Vitae, and References upon Request". Once contacted by a prospective employer, you can send this information along with a cover letter introducing yourself and stating that you look forward to the interview. NOTE: FUTURE RESUME BOOKLETS WILL HAVE A 1 PAGE LIMIT. A SECOND PAGE WILL BE ALLOWED AT A COST THAT IS YET TO BE DETERMINED. NO RESUMES OVER 2 PAGES WILL BE ACCEPTED. Taking the suggestion supplied by the many reviewers a format will be adopted. Although the format may be altered slightly to better emphasize one's experience, any resume that greatly deviates from the standard will not be accepted. If any Governor finds the standard format lacking for their particular country, additional formats may be adopted. The standard format is currently being adopted and will be published in the next newsletter, along with a request for resumes. No resumes are currently being accepted. Another added point that needs to be made is that many resumes were received past the deadline and did not get into the booklet. Future deadlines will continue to be honored in the same manner. ----------------------------------------------------------------------------- | Undergraduate Issues -- Submissions to: | ----------------------------------------------------------------------------- by Thomas Edwards This "Undergraduate Issues" column is a way of showing our commitment to undergraduates, and to facilitate their inclusion in a field dominated by graduate students and professionals. Although we have not received any submissions for this column since the first issue, in the future we hope to present undergraduate research opportunities, highlight papers written by undergraduates, and present other material to complement the general-interest topics that appear in other columns. In our next issue, this column will feature a review by undergraduate Thomeas Edwards of the new concept of "Compostional Learning". What appears here, though, is to a large part up to you. Are you an undergraduate involved in neural network research? Do you have neural net research/programming positions available for undergraduates? If so, send us a note by electronic or surface mail! ----------------------------------------------------------------------------- || Academic Programs -- Submissions to: || ----------------------------------------------------------------------------- by Andrew Worth The format of the column in this issue will be different than usual. We are endeavoring to build an academic program database that will be useful to the entire research community. Towards this goal, we had people fill out "Academic Program Questionnaires" at the two big neural network conferences this summer. The results are presented below. The next ISSNNet newsletter should see this column return to its normal format of presenting descriptions of academic programs by ISSNNet members. Over 30 Schools responded from 11 countries: West Germany (5), South Africa (1), Italy (3), France (6), Belgium (2), Switzerland (2), Cyprus (1), England (3), Canada (1), the Netherlands (1), and the USA (12). The following schools are accepting students in neural networks: Approx. accepting students: School Size Ugrad Grad Postdoc ............................................................................. Arizona State Univ. (USA) x Boston University (USA) 70 x x Catholic Univ. (USA) x Columbia University (USA) x x Cyprus Inst. of Neurology and Genetics (CYPRUS) x x Ecole Polytechnique Federale (FRANCE) x x x Ecole des Mina de Paris (FRANCE) x x Free Univ. of Brussels (BELGIUM) 2+ x IEF Obsay Paris SUD (FRANCE) x x Inst. Nat. Polytechnique Lorraine (FRANCE) 4+ x Inst. Nat. Polytechnique de Grenoble (FRANCE) x Kath Univ Leuven (BELGIUM) 7 x x Stanford University (USA) x Technical University Munich (GERMANY) 4+ x x Univ. of Arizona (USA) x x Univ. of Toronto (CANADA) x Universita Degli Studi Di Napoli (ITALY) 7 x x University of Berne (SWITZERLAND) 3+ x x x University of Bonn (GERMANY) x x x University of California, San Diego (USA) 60 x x x University of Cincinnati (USA) x x University of Hawaii (USA) 6 x x University of Rome (ITALY) 19 x University of Sussex (UNITED KINGDOM) 9 x x University of Texas (USA) 14 x University of Twente (THE NETHERLANDS) 20 (short term exch. all levels) University of the Witwatersrand (S. AFRICA) 5 x _____________________________________________________________________________ ISSNNet Request for Academic Program Information The International Student Society for Neural Networks (ISSNNet) is compiling a list of academic programs relating to Neural Networks from around the world. We would like your input if you are a member of a scholastic program that is in any way related to Neural Networks, Parallel Distributed Processing, Connectionism, Computational Neuroscience, Neural Modeling, Neural Computing, etc. We hope to provide this service so that (1) interested students will be able to apply to those programs that will most closely satisfy their educational goals, and (2) current students and non-students will be aware of existing academic programs. This service is intended to not only provide an overview of these programs and contact points for more information, but also a personal glimpse into what's behind the official descriptions. your submission will be taken as permission to make the information public unless we are otherwise notified. Information should include: - Official address to contact for more information (surface mail and email) - Official description of the program. - Names of Faculty Members and their interests - Degrees requirements (BA, BS, MA, MS, PhD, etc.) - Short description of courses offered - Computing resources (Hardware and Software Tools) - Number of Students (grad/undergrad) and related faculty - A brief *personal* description of the program, department, etc. describing motivation, emphasis, goals, and/or overall ambiance. - Student Contacts (w/ telephone numbers, email and surface addresses, degree sought, interests, and date of graduation) This information is above and beyond the academic questionnaires that were filled out at the San Diego and Paris conferences and will be available to members by electronic mail. In the future, an ftp site may also be avaliable. Members without access to electronic mail should contact us in writing, or contact their local Governor. Coordinated responses from each institution are encouraged and will be appreciated. Please submit descriptions of academic programs in plain text by electronic or surface mail. ----------------------------------------------------------------------------- |-| EDITORIALS -- Submissions to |-| ----------------------------------------------------------------------------- Editor's note: we are always glad to receive feedback on the newsletter, whether it is compliments or complaints. Some of the editorials will be published in subsequent issues. Rohan Baxter writes: I enjoyed the first edition of the Newsletter. Its great that ISSNNet is so international. May I suggest that "Spring Edition" (or any other season) is an inappropriate designation for such a world-wide journal. Its autumn in the Southern Hemisphere- I can feel the chill in the air! However basically we do manage to keep the same calender down here- so that may be a more general means of naming your editions. Keep up the good work, Rohan. Editor's reply: Well, I'll never cease to be amazed by the ease with which we forget such basic things. Hopefully we won't make the same faux pas on the next issue. _____________________________________________________________________________ John Reynolds writes: I wanted to let you know I enjoyed reading the ISSNNet newsletter. It could potentially provide a platform on which to build unity in the field, without the political intrigue some of our elders have been involved in. If there is some way I can help out, please let me know. A couple of observations: Especially considering this is the first issue, I think it was in bad taste to characterize work in back-propagation as "attention grabbing." This implies "flashy and without substance" to my mind. Second, while it may indeed be valid to argue in favor enhancing the interdisciplinary nature of the field, your segregation of neural network research groups into "those of us that are lucky enough to belong to such (an interdisciplinary) program" and those who need our help was, I fear, a bit of a snub. While we certainly owe it to the field to promote what we perceive to be the proper approach to research, it will be impossible to do so if we antagonize the vast majority. Other than that I was, on the whole, pleased and impressed, and send my congratulations and encouragement to you and all those who were involved in publishing the first issue. Paolo Gaudiano replies: Thanks for the note. I have heard exactly the same remarks from a couple other people, so I suspect other readers may have felt the same way. The backprop remark was not meant to sound as pejorative as some may interpret it, but rather to suggest that some of the work that put backprop in the spotlight was meant to excite people. In particular, I personally found the NetTalk demo to be an incredibly compelling demo, much more than someone telling me "backprop can do nonlinearly-separable maps", although I now feel that the mathematical properties of backprop are more important than its ability to learn mappings between specially-constructed data sets. Perhaps "captivating" would have been a better term. About your second remark, my comment reflects the subjective opinion that the study of NNets AS A FIELD requires an interdisciplinary approach, something only a few students have the opportunity to do. This does not imply that anyone who only wishes to study a particular application of neural nets is any less of a researcher. Rather, the comment was motivated by the observation that there are many students who would like to learn more aspects of neural nets, but they simply do not have the opportunity to do it through their academic institution. If other students can provide additional sources of information through ISSNNet, then everyone will benefit. ----------------------------------------------------------------------------- ||-|| Action Potentials: ISSNNet Announcements || ----------------------------------------------------------------------------- Editor's Note: Please submit all SHORT announcements to the address above, or send surface mail to our main address (front page). Call for help: be a part of the ISSNNet organization! ---------------------------------------------------- As the society is growing, we are finding out that it takes more than a handful of people to take care of it. Part of the effort in writing the official ISSNNet bylaws was the description of positions on the governing board, to insure that the workload would be adequately distributed among a number of people, so that the well-being of the society would not depend on few individuals. This is particularly important for a student society, as we all know that the amount of spare time of a student is highly variable, and of course we do not want society matters to interfere with good scholarly conduct. The bylaws call for a flexible governing board, consisting of four society Officers (President, Vice President, Director, and Treasurer), and a number of "Governors." The number of Governors on the Board at any time depends on the number of members: any group can select a Governor by sending us a nomination letter with the name and signatures of at least ten active members. The Governor will primarily act as a liaison between the society and the local members. As we try to provide information exchange through various means of communication, it becomes important to optimize the distribution of information. We are charging a $5 annual fee, and it can cost $2 to send a single newsletter by surface mail to foreign countries. Even electronic mail can be costly across continents. With a local Governor, it becomes possible to send a single copy of a newsletter (in various formats) to one or few locations, and have it circulated locally at a much lower cost. The Governors should also keep the society informed of local events, maintain a list of local members, and organize annual membership drives. We already have a number of Governors for various countries around the world, but we need more, particularly within the USA. Because of the multi-media nature of ISSNNet, it is important that Governors have access to an account with electronic mail on a major network (e.g., ARPANET, JANET, BITNET, etc.). If you can volunteer an average of one or two hours a week (flexible), send electronic mail to to receive further details. The four Officer positions will be filled by election. We are trying to organize the first elections before the end of the year. The main duties of the Officers will be to represent ISSNNet and to insure that all vital society matters are taken care of. This means primarily to oversee volunteers assigned to different tasks. If you are interested in running for one of the officer positions, send us a message for more details . The following is a list of some of the tasks for which we need volunteers, if you wish to help with one of these, send an e-mail message to: . *) Maintain overall membership directory. *) Organize yearly elections. *) Organize membership drive. *) Maintain our Industry directory. *) Maintain Graduate School Directory. *) Maintain Computer database. *) Maintain electronic mailing lists. *) Basic accounting. *) Organize fundraising events. *) Organize meetings at conferences. *) Oversee/coordinate all governors. *) Put together the ISSNNet newsletter. All of these positions are on a voluntary basis, although we are hoping to set up small "scholarships" as a way of supporting students that help with these and other tasks. This is primarily an opportunity for students to get involved with ISSNNet, and can be a good way to gain experience and meet people. Most of these positions require self-organizing people that can work creatively and independently, because many of these tasks need to be done for the first time, and there may not be anyone to provide exact details of what to do. We hope to be able to provide general guidelines and the name of a person to contact for general questions and assistance. Finally, a number of people have suggested useful projects that would be ideal as ISSNNet activities. If you have any projects that you'd like to organize, let us know and we'll try to work something out. Some examples would be organizing special meetings, or setting up a database of public-domain neural net simulators. Again, you should send ideas to: . =================== We need a computer! =================== We are still looking to get someone to give us access to a computer on one of the major networks that we can set up for the following tasks: *) To maintain our e-mail list and to use as a centralized mail server. *) To set up an archive for a USENET newsgroup that will complement the electronic mailing list. *) To set up an anonymous FTP site for access to various ISSNNet information (academic programs, articles, software, whatever). *) To take care of general ISSNNet tasks, such as editing and formatting the newsletter, maintaining the membership list, etc. The ideal solution would be a UNIX workstation with telnet access, on which we could set up an account with administrative privileges. We have a number of experienced students that would be willing and able to set up the ftp site and the other things. Because we are offically incorporated as a non-profit organization, equipment donations can be used as a tax write-off. Alternatively, if an instituion were willing to let us have control of a portion of an existing anonymous FTP site, and maybe an account with a few MB of disk, that would be a good start. If you can help, please contact us by electronic or surface mail (addresses on front page). ----------------------------------------------------------------------------- SPONSORS: We are now an official non-profit organization. We need monetary and material support. All our profits are used to support students (with minimal overhead through much volunteer work). We would like to receive travel and registration money for students presenting papers at conferences. If you want to sponsor an individual paper, or wish to contribute in other ways, please send us electronic or surface mail (addresses on front page). CONFERENCE UPDATE: ----------------- We wish to thank again the organizers of IJCNN and INNC for donating booths to ISSNNet. Look for us at these upcoming conferences: IEEE Conference on Neural Information Processing Systems (NIPS). Denver, Colorado (USA). November 26 - November 29, 1990. Workshops through December 1, 1990. For information contact Kathie Hibbard, NIPS*90 Local Committee, Engineering Center, University of Colorado, Campus Box 425, Boulder, CO 80309-0425, USA. Second Australian Conference on Neural Networks (ACNN). The University of Sidney, Sidney, Australia. February 4th-6th, 1991. For information contact Ms. Justine Doherty, Secretariat ACNN'91, Sidney University, Electrical Engineering, NSW 2006 Australia. Tel: (+61-2) 692-3659; Fax: (+61-2)692-3847; Email: . NNet CONFERENCE ORGANIZERS: Show your support to students! Offer low student rates or help our Society. Send e-mail to . ----------------------------------------------------------------------------- --- ISSNNet T-shirts Now Available --- We had a very nice T-shirt made at the beginning of summer. The shirt is 100% cotton, light blue, and it features a colored ISSNNet logo. They were a great success at the IJCNN and INNC conferences this summer, and now you can purchase one by mail. The price is $12.00 per shirt for ISSNNet members and $14.00 for non-members, plus $1.00 per shirt for shipping costs. We will try to coordinate the T-shirt sales with our Governors for countries outside the USA, and the price may vary slightly. In either case, please allow 6-8 weeks for delivery. All orders must be pre-paid by check or money order. Fill out the following information and mail it with your payment to our main address (front page): Name:__________________________________ Address:__________________________________________ __________________________________________ __________________________________________ __________________________________________ Size/Quantity: Small___ Medium___ Large___ X-Large___ Total T-shirts ordered:____ x $13.00 ($15.00 for non-members) = ___________ Make check or money order payable to "ISSNNet". Non-US members: if you do not have a Governor to contact, send e-mail to . >>>>>>>>>>>>>>>>>>>>> BECOME AN ISSNNet MEMBER !!!!! <<<<<<<<<<<<<<<<<<<<<<<< The single most important way to show your support for ISSNNet is to fill out the membership form below. Your money is used in part to defray the cost of printing this newsletter for distribution at conferences, or to support other ISSNNet programs. All profits go to student support. Membership includes a subscription to this newsletter. There are several types of membership, all cost (US)$5. Students, prospective students, and recent graduates are eligible for regular memberships. Only regular members may vote or hold office. Faculty and staff of recognized academic institutions may join as academic members. All others may become affiliate members. ISSNNet bylaws and other details are available upon request . NOTE FOR NON-USA Members: if your country has a Governor, his or her name should appear on the membership form. Make your payment to that person's name in your local currency. The fee will be the round sum nearest $5 in your currency, as decided by your Governor. ---- MEMBERSHIP LIST UPDATE: ---- We apologize to all the members who have sent in their fee and never heard from us. We are compiling a full membership list to be sent out to all members. However, we have a lot of incomplete information for many of you. Unfortunately, most of the forms we sent out did not ask for the member's interests in NNets, which is one of the main purposes of a membership list. We ask that everyone send a brief message to describing what aspects of NNets you are interested in. Cut here ----------------------------------------------------------- Cut here |||| 1990 ISSNNet Membership Form |||---------------------------------|||| We encourage all friends of neural networks to join the society. Please send this completed form and a check for $5 (in US dollars) to ISSNNet. This is a one-year membership that expires June, 1991. Name (First Last):_________________________________ Address:______________________________ Mail to: ISSNNet Memberships ______________________________ P.O.Box 557, New Town Br. ______________________________ Boston, MA 02258 ______________________________ USA School/Company:____________________________________________ e-mail (include network) :_________________________________ (PLEASE PRINT CLEARLY TO AVOID LOST MAIL!) Your Interest in NNets:______________________________________________________ _____________________________________________________________________________ Membership Type: Regular/Student( ) Academic( ) Affiliate( ) 1. May we include this information in our membership directory? yes( ) no( ) 2. May we put you on other mailing list to fund the society? yes( ) no( ) 3. To help defray printing and mailing costs, may we send you electronic mail only? yes( ) no( ) 3a.If you marked "yes" to question 3, what formats do you prefer? PostScript( ) LaTeX( ) Plain text( ) signature:____________________________ date:_________________________