I participated in CEDEC 2017

I participated CEDEC 2017 from August 30th to September 1st in Yokohama. The conference is the first time I went to DECEC that is held every year as far as I know. All participants engage in game developing.

I look at below sessions.

Day 1

Shadowverseにおける「デッキのトレンド分析」を題材としたデータマイニング技術の活用手法紹介

They have tried to analyze card deck’s trends that user in the game use. They use a technique like “edit distance” among different decks.

アイドルマスター シンデレラガールズ ビューイングレボリューション制作事例 ~最高のVRライブ体験に必要となる要素とは~

They created the great experience of a live concert in VR. It is important to render mob’s detail whom body, showing the light on a body it reflected by lighting sticks in VR and mob’s cheers.


Day 2

- 人工知能(メタAI)を用いたゲームデザインの変革
[人工知能学会×CEDECコラボセッション] 人工知能と対話システム ~キャラクターとの自然な会話~
データ分析の固定観念を覆す -ユーザー体験を向上し続けるモンスターハンター エクスプロアのデータサイエンス-

Day 3

遺伝的アルゴリズムによる人工知能を用いたゲームバランス調整

They have been creating the player AI that emulates a game player. The reason to create it that reduce quality assurance costs because a lot of contents in a game are created continuously, and they need to check contents also created continuously. It is easy we could imagine that the contents created which increase the cost of quality check.

キャラクターらしさ学習AI: 多数のキャラクターの個性や違いの可視化によるシナリオライティング支援システム事例

This support a writer in writing scenario for games with a script written by a writer.
The AI learns characteristics of a character in a game and classifies a script by it into several characters. A writer can fix the script with the scores calculated by the AI. The score indicates a likelihood of characters in a game is like this:

Character Name Score
A 10
B 5
C 2

This example indicates that a script might be the A because its core is higher than the others.

クラウドの分析システムを活用したアクセスログにもとづくリアルタイム不正ユーザBANシステム

They have built the system that bans the user request illegal accesses to game API servers in real time.

Leave a Reply

Your email address will not be published. Required fields are marked *