プロンプト:food photography, hamburger in the style of a mcdonald’s big mac, but luxurious michelin kitchen style –v 5
プロンプトを使ってハンバーガーの画像を生成してもらった例
プロンプト:80s lotus esprit chrome details, nostalgic and timeless, warm and soft, sepia tones, deserted industrial warehouse, Fujifilm X – T4 with 35mm lens, shallow depth of field, early morning light
プロンプトを使ってスポーツカーの画像を生成してもらった例
プロンプト:Joe biden and donald trump laughing and playing golf together, mid day, natural direct sunlight, photography, eye level shot
プロンプト:food photography, hamburger in the style of a mcdonald’s big mac, but luxurious michelin kitchen style –v 5
プロンプトを使ってハンバーガーの画像を生成してもらった例
プロンプト:80s lotus esprit chrome details, nostalgic and timeless, warm and soft, sepia tones, deserted industrial warehouse, Fujifilm X – T4 with 35mm lens, shallow depth of field, early morning light
プロンプトを使ってスポーツカーの画像を生成してもらった例
プロンプト:Joe biden and donald trump laughing and playing golf together, mid day, natural direct sunlight, photography, eye level shot
具体的には、主語の性別を格納するメタデータを文に追加する方法。例えば、”You are very nice “という文は英語では性別が曖昧だが、並行するポルトガル語の文が “Tu és muito simpática “であれば、英語の文の最初にタグを追加して、モデルが正しい翻訳を学習できるようにする。
Here’s a sentence describing what Google’s home page should look and here’s GPT-3 generating the code for it nearly perfectly. pic.twitter.com/m49hoKiEpR
Q: 1955年のアメリカ大統領は?( Who was president of the United States in 1955?) A: 1955年のアメリカ大統領はドワイト・D・アイゼンハワーです(Dwight D. Eisenhower was president of the United States in 1955.)/i>
Q: 彼はどの政党に所属していたか?(What party did he belong to?) A: 彼は共和党に所属していました(He belonged to the Republican Party.)
次に常識に関する質問だ。従来のAIにとっては難問であったが、GPT-3は回答出来ている。
Q: 大好きな動物は?(What is your favorite animal?) A: 犬がお気に入りです(My favorite animal is a dog.)
Q: どうして?(Why?) A: 犬は忠実で友好的だからです(Because dogs are loyal and friendly)
Q: 犬が不機嫌になる2つの理由は何?What are two reasons that a dog might be in a bad mood? A: 犬が不機嫌になる2つの理由はお腹がすいているか暑い時です(Two reasons that a dog might be in a bad mood are if it is hungry or if it is hot.)
Q: 1600年のアメリカ大統領は?(Who was president of the United States in 1600?) A: 1600年のアメリカ大統領はエリザベス1世です(Queen Elizabeth I was president of the United States in 1600.)
いくつかの関連した質問をした場合も、前後関係を理解した回答ができていないという。
Q: 1本鉛筆を箱に入れ、さらにもう1本鉛筆を箱に入れたら箱の中には何がある?(If I put a pencil in a box, then put another pencil in the box, what is in the box?) A: 鉛筆が2本(Two pencils.)
Q: そこから1本鉛筆を取り出して、カエルを入れたら箱の中に何がある?(Then if I take out a pencil and put in a frog, what is in the box?) A:カエルが1匹( A frog.)
#gpt3 is surprising and creative but it’s also unsafe due to harmful biases. Prompted to write tweets from one word – Jews, black, women, holocaust – it came up with these (https://t.co/G5POcerE1h). We need more progress on #ResponsibleAI before putting NLG models in production. pic.twitter.com/FAscgUr5Hh
Open AIで代表を務めるSam Altman氏も、GPT-3はまだ問題が多くありGPT-3はAI時代の兆しでしかないとツイートしている。
The GPT-3 hype is way too much. It’s impressive (thanks for the nice compliments!) but it still has serious weaknesses and sometimes makes very silly mistakes. AI is going to change the world, but GPT-3 is just a very early glimpse. We have a lot still to figure out.
グローバルに求められるサービスを、「どこで」「誰と」「どのように」作り上げていくか。2019年11月5日に開催された「DESIGN for Innovation 2019」での『AI Startupが挑戦するグローバルで通用するサービスデザインとは』のセッションでは、株式会社シナモンの平野未来氏とビースポーク株式会社の綱川明美氏からお話を伺った。
事業としては、 非構造データをアクションにつながるインサイトへと転換していくAIの構築を行っています。非構造データは、メールや画像、音声やPDF、またワードやパワポなどのファイルに載っている情報ほとんどを指します。このような IT 時代なので、各社に大量の情報があるのですが、うまく活用されずに眠っている現状があります。
波江(btrax, Inc. モデレーター):
お二人とも、グローバルな市場にサービスを展開されていますが、平野さんは、日本企業をターゲッティングされていましたし、綱川さんも訪日外国人向けにサービスをスタートされていましたよね。グローバル展開を狙うサービスを、日本拠点にスタートすることについてどう考えていらっしゃるか教えてください。
今回は自動車産業、音声認識、新型iPhone(AR)、人工知能、ソーシャルメディアなど一見マーケティングに関係ないように思える最新技術のトレンドを列挙した。しかし、これらは今後マーケティング活動に大きな影響を与える可能性がある。今までの成功事例や経験を元に作り出すマーケティング施策の実行も大事だが、投資家が最新技術に出資をするようにマーケターも新しい技術を意識したチャレンジングな投資的マーケティングも必要なのかもしれない。
参考:18 Marketing Trends to Watch in 2018