| 13:30--14:30 | 大隅剛史（岩間研） | Enumerating Isolated Subgraphs |
| 14:40--15:40 | 原口和也（永持研） | Classifiers Based on Iterative Compositions of Features |
| 15:50--16:50 | 杉山和徳（[株]ソフィア・クレイドル 代表取締役社長） | ベンチャービジネスの立ち上げ方 |
リンクの張ってある題名は、クリックすると概要が見られます。
終了後、田村先生の送別会を行います。
場所：アラビアンロック ->[詳細]
京都市中京区新京極四条上ル中之町583-2 ミマツワールド1F
075-241-7822
時間：18:30〜
申込はこちらのページからお願いします->[申込のページ]（申し込み締切り：９月１６日(木)）
- One of the recently popular problems on the web graph is to find
- groups of web pages which are densely linked inside each group.
- Cliques are obviously a natural model for such groups.
- Unfortunately, however, most problems regarding cliques, such as finding a
- maximum one and enumerating maximal ones, are intractable.
- There is no good approximation, either.
- In this paper, we introduce ``isolated'' cliques or isolated subgraphs in
- general, for which we require not only that each subgraph is densely linked inside
- but also that each subgraph has few links to the outside.
- Typically, if the subgraph is of size $k$,
- then only $ck$ ($c$ is a constant) or less edges to the outside are allowed.
- Our main results show that such an isolation condition makes
- many problems surprisingly easy.
- Problems like enumerating isolated cliques and isolated pseudo cliques
- (which are less dense than cliques) can be solved in polynomial time.
- −
杉山和徳（[株]ソフィア・クレイドル 代表取締役社長）[略歴]