Famous conferences on AI and machine learning

Transfer from Forum http://www.ieee.org.cn/dispbbs.asp? Boardid = 62 & replyid = 31567 & id = 29962 & star = 1 & Skin = 0
The author seems to be Miss Zhou Zhihua from the Southern University.

Several AI conferences I know (first-class)
The following parts are listed in alphabetical order:

Ijcai (1 +): the best comprehensive meeting of AI. It will be held every two years starting in 1969 and will be held every odd year. because AI is too big, although each session can basically record more than 100 (now more than 200), there are only a few articles in each field, there are about 10 articles every time in fields such as machine learning and computer vision, which makes it very difficult. however, from the perspective of the recording rate, it is not very low, basically about 20%, because the number of pedestrians in the system is too small. If there is no hope, do not waste the reviewer time. recently, articles sent from Chinese mainland to international conferences are similar to the tide, and because there are very few research groups in China that can check their own, therefore, many meetings are in complain, saying that low quality articles in China severely impede the efficiency of PC work. in this case, it is estimated that the recording rate of international conferences will fall over the past few years. in addition, the previous ijcai did not have a poster. In the past 03, two pages of poster were added to reduce the number of good people who were mistakenly killed. it is worth mentioning that ijcai is a company that looks like "ijcai Inc. "sponsored by (not actually a company, but actually a Foundation), several Awards will be presented at each meeting. The two most important ones are ijcai Research Excellence Award and computer & thoughts award, the former is a Lifetime Achievement Award, and each person is basically the highest award of AI (interestingly, two of the six winners of the AI-based awards have not yet received this award ), the latter is awarded to young scientists under 35 years of age. the speeches of these two awards were a major event of ijcai each time. in addition, the PC member of ijcai is equivalent to the area Chair of other meetings and has great power, because it is reviewed by the PC member for reviewer, and unlike the PC member of general meetings, it is actually a reviewer. in order to restrict such power, the Cai reviewer assigns two PC member members for each article, and the primary PC member finds three reviewer members and the second PC member finds one.

Aaai (1): Annual Meeting of the American AI society aaai. it is a good meeting, but its grade is unstable. It can be 1 + or 1-or 2 +. In general, I will give it "1 ". this is because its launch method is completely restricted by ijcai: annual launch, but if ijcai is held in North America this year, it will be suspended. therefore, in even years, because ijcai is not available, it is the best comprehensive AI conference, but because of its appeal, after all, it is smaller than ijcai, in particular, Europeans hold the aaai field much less than ijcai (in fact, the Asians are also), so it is a little weaker than ijcai, basically between 1 and 1 +; in an odd year, if ijcai is not in North America, aaai will naturally become a level-1 meeting (1-or 2 +) lower than ijcai. For example, in 2005, there will be both ijcai and aaai, and the two meetings will be coordinated, this makes the time for the Cai recruitment notification a few days earlier than the aaai deadline, so that the articles selected by ijcai can be sent to aaai. during the review, the PC Chair of ijcai has been urging everyone to speed up, Because aaai has been worried that the recruitment notice of ijcai will be delayed and aaai will be in trouble.

Colt (1): This is the best computing Learning Theory conference, sponsored by ACM and held every year. computing learning theory can basically be seen as the intersection of Theoretical Computer Science and machine learning, so this will be seen by some people as a theoretical computer science meeting rather than an AI meeting. A friend of mine described it in one sentence: "a small group of mathematicians are in a meeting ". because the colt field is relatively small, the annual meeting is basically those people. one interesting thing to mention here, because there have been too many conferences in China recently, and many of them are from lncs/lnai, and lncs/lnai have basically been corrupted, however, unfortunately, lncs/lnai have some good meetings, such as Colt.

Cvpr (1): One of the best conferences on computer vision and pattern recognition, sponsored by IEEE, held every year. although there is computer vision on the subject, I think its pattern recognition tastes heavier. in fact, it should be the best meeting for pattern recognition, while in computer vision, there is also an equivalent of iccv. IEEE has always had a tendency to make it a "grand event", and some of them will be well organized as "grand events" from quality in history. this is the way to do cvpr. many articles have been recorded over the past few years. the Chair responsible For the cvpr meeting recently sent a message saying that for this community, it is worse to let good people be killed by mistake than to be missed by bad people. So should we reduce the chances of killing good people by mistake? So I estimate that cvpr will expand in the next year or the next year.

Iccv (1): As mentioned in introducing cvpr, one of the best in computer vision. sponsored by IEEE. iccv was held in an odd year. It used to take turns in North America, Europe, and Asia. It was originally set in Beijing in 2003. Later, it was changed for SARS and France, which was originally set to 05 years. Iccv '07 will be held in South America (Brazil) for the first time.
In principle, cvpr runs in North America every year. If iccv is in North America that year, there is no cvpr in that year.

Icml (1): One of the best conferences on machine learning. It is now hosted by imls and held every year. See introduction to nips.

NIPS (1): One of the best conferences on Neural computing, sponsored by nips, is held every year. it is worth noting that this meeting will be held in the same place every year. It used to be Denver in the United States and is now in Vancouver, Canada. In addition, it will be held at the end of the year, and the proceedings will be published only 2nd years after the Conference is complete, that is to say, Nips '05 is published in. the meeting name is "Advances in neural inxxxxation processing systems". Therefore, unlike "standard" Machine Learning meetings such as icmlecrf, Nips contains a considerable amount of neuroscience content, and machine learning. however, since the main body of the Meeting is machine learning, or is closely related to machine learning, many people regard nips as one of the best meetings in machine learning. this meeting is basically controlled by the disciples of MichaelJordan. Therefore, it is not difficult for the Jordan team to send nips, and some work that may not be very strong can also be done, however, it is very difficult for people outside this circle to publish an article, because it is very small for "outsiders. therefore, it is more difficult to issue nips than icml for users other than Jordan. in other words, icml is relatively open, and the influence of small circles is not as big as that of NIPs, So North America and Europeans recognize it, while some nips (especially some Europeans, including some people) refuse to contribute. this is not a good thing for the Conference itself, but it does not seem to care much because Jordan is very powerful. recently imls (International Machine Learning Society) was elected to the Council. Eligible candidates include those who have posted articles in icmlecmlcolt in the past three years, and Nips are excluded. in any case, this is a very good meeting.

ACL (1-): The best conference in computational linguistics/Natural Language Processing, ACL (Association

Sponsored by computational linguistics.

KR (1-): One of the best conferences in terms of knowledge representation and reasoning is actually one of the best conferences in traditional AI (that is, logic-based AI. KR Inc. host, now an even number of years.

SIGIR (1-): The best conference on information retrieval, sponsored by ACM, is held every year. this will be getting heavier and heavier now. information retrieval should not be considered as AI, but because machine learning is used more and more, it has become a bit of a taste of machine learning application meetings in recent years, so it is also listed.

Sigkdd (1-): The best Conference on Data Mining, sponsored by ACM and held every year. this Conference has a short history. After all, compared with other fields, data mining is just a little brother or even a little nephew. it was difficult to include it in tier-1 a few years ago. On the one hand, it was far less famous than other top Conference, and on the other hand it was relatively easy to be hired. but now it is undoubtedly listed in tier-1. in the past few years, KDD has a high quality. since 2000, the recording rate of sigkdd full paper was between 10% and 12%, far lower than that of ijcai and icml.

It is often said that KDD is more difficult than ijicai and icml. Ijicai has only 6 pages, while KDD has 10 pages. Without a solid system, it is difficult to leave vulnerabilities. Many ICAI customers also invest in KDD every year.

UAI (1-): its name is "uncertainty in artificial intelligence". It involves many aspects such as reasoning and learning. It is hosted by auai every year.

Several AI conferences I know (second, third stream)
(Originally Daniel on lilybbs. US)

This is my opinion. It is for reference only. Tier-1 is fully listed, tier-2 is incomplete, and tier-3 is incomplete.
Sort the same score in alphabetical order. not strictly speaking, tier-1 can be enviable, and tier-2 can be respectable because there are many AI-related meetings, therefore, it is good to list data in tier-3.

Tier 2: Tier-2 meetings are not fully listed. I am familiar with more fields.

Aamas (2 +): The best meeting on the agent. But now the agent is a general concept,
Almost all AI-related conferences have such content, so the trend of aamas decline is very obvious.

Eccv (2 +): second only to iccv conferences in computer vision, because this field is growing fast and may be upgraded to 1-bit.

Ecml (2 +): The Conference second only to icml in terms of machine learning. Europeans strongly join us. Some people think it is already 1. I keep it a little bit and still put it in 2 +. as Machine Learning develops rapidly, the reputation of this conference has increased significantly.

ICDM (2 +): second only to sigkdd in Data Mining. Currently, it is equivalent to SDM. this will have only five years of history, and the rising speed will be amazing. A few years ago, ICDM was far behind pakdd, and now it is very far away.

SDM (2 +): second only to sigkdd in data mining, which is equivalent to ICDM currently. siam has a thick base, but the impact in CS is smaller than that in ACM and IEEE. SDM is expected to be exceeded by ICDM, but at least it is still equivalent.

Icaps (2): The best conference on artificial intelligence planning was merged from the previous international and European planning conferences because the field has gradually become increasingly deserted and the impact has been lower than before.

Iccbr (2): The best meeting in case-based reasoning. Because the field is not too big and it has been half cold and hot, it always stays on 2.

Colling (2): computing linguistics/Natural Language Processing is second only to ACL, but the gap with ACL is much larger than ICCV-ECCV and ICML-ECML.

Ecai (2): It has been a long history of the Integrated Artificial Intelligence Conference in Europe. However, due to the pressure of ijcai/aaai,
It's hard to go up.

Alt (2-): It's a bit like Colt's tier-2 version, but because there aren't many people engaged in computing learning theory, the group is just a little better, basically it's up to Colt, so there are a lot of things in ALT that are not computing learning theory.

Emnlp (2-): a good meeting in computational linguistics/Natural Language Processing. Some people think it is equivalent to Colling, but I think it should be weaker.

ILP (2-): The best meeting in logic program design, but because many other meetings have ILP content, it can only hold the 2-position.

Pkdd (2-): Data Mining Conference in Europe, which currently ranks 4th. the European would like to raise it up, so it has been bundled with ecml over the years, hoping to use ecml to bring it up. but because of ICDM and SDM, this is no longer possible. therefore, this year's pkdd and ecml are still open together, but they have already been reviewed independently. (In the past, we were able to submit two meetings at the same time. The author can declare which one would give priority, if ecml is not available, it can be accepted by pkdd ).

Tier 3

Accv (3 +): The Asia-Pacific Computer Vision Conference is a good example.

DS (3 +): A Japanese Conference on Data Mining.

Ecir (3 +): European information retrieval conference, which was just a UK information retrieval conference a few years ago.

Ictai (3 +): IEEE's most important artificial intelligence conference, which is applicable to applications, is a typical example of IEEE. in the past, quality was good, but the longer it took, the worse its reputation was. The worse it was, the worse it seemed that it was still declining. In fact, 3 + is not too good.

Pakdd (3 +): Asia Pacific Data Mining Conference, which currently ranks 5th.

ICANN (3 +): the European Neural Network Conference is the best in terms of quality. However, people in this field do not pay attention to the Conference. In this field, it is not as important as ijcnn.

Ajcai (3): Australia's Integrated Artificial Intelligence Conference, which is good at national/regional AI conferences.

CAI (3): Canada's Integrated Artificial Intelligence Conference was a good one in national/regional AI conferences.

CEC (3): One of the most important sessions in evolutionary computing, grand event. ijcnn/CEC/FUZZ-IEEE these three meetings are the most important aspects of Computing Intelligence or soft computing, they often open together, then called wcci (World Congress on Computational Intelligence ). however, this field is not the same as other branches of CS. It is similar to other disciplines. It only attaches importance to journal and conferences, so the recording rate is usually around 85%, there are also exercises for beginners.

FUZZ-IEEE (3): Fuzzy aspects of the most important conference, grand event, see the CEC introduction.

Gecco (3): One of the most important sessions in evolutionary computing, equivalent to CEC, grand event.

Icassp (3): One of the most important speech conferences. People in this field are not very concerned about the Conference.

ICIP (3): One of the most famous conferences in image processing, event-oriented.

ICPR (3): One of the most famous conferences in pattern recognition, event-oriented.

IEA/AIE (3): AI Application conference. generally, only a few articles are used to nominate excellent papers at meetings. It is already a high honor to be nominated. This will be very interesting. Every time I nominate 1 or 20 excellent papers, it is very lively to have several sessions dedicated for the report of the nominated paper.

Ijcnn (3): The most important meeting on neural networks. For more information, see the CEC introduction.

Ijnlp (3): a well-known Conference on Computational Linguistics/natural language processing.

Pricai (3): Although the history of the Asia-Pacific Integrated Artificial Intelligence Conference is not short, it is difficult to rise because it is better than it or a considerable number of integrated conferences.

Famous conferences on AI and machine learning