🚧 Under construction 🚧
Welcome to the Language Media Laboratory
At the forefront of AI innovation, our lab merges creativity with functionality to redefine how machines comprehend and utilize natural language. We are a pioneering force in the AI landscape, driven by a multidisciplinary approach that transcends classical NLP. Our research combines insights from diverse fields with both the latest and historical advancements in AI, offering unique solutions and breakthroughs.
Our Research Focuses:
• Multidisciplinary Approach: We integrate knowledge from various disciplines, pushing the boundaries of traditional NLP to discover new, impactful ways of understanding language.
• Dialog Processing: We enhance the naturalness of dialog by addressing not only semantic aspects but also pragmatic ones. From prosodic features to user modeling, our work covers a comprehensive range of dialog-related topics.
• Language and Knowledge Acquisition: Exploring how machines can effectively acquire and comprehend knowledge, we experiment with semantic primes and fluid context management to achieve more human-like understanding and learning.
• Affective Computing: Emotions are integral to human interaction. Our work on emotional understanding in AI, for example through our ML-ASK system, has been a cornerstone since 2001, influencing areas like machine ethics, language acquisition, and humor processing.
• Moral Machines: Since 2005, we’ve developed ethical decision-making algorithms inspired by classic ethical theories, including Kohlberg’s stages of moral development, utilitarianism and deontology, harnessing the Wisdom of the Crowd.
• Humorous and Figurative Language: Language creativity, such as metaphors, poetry, and humor, still poses challenges for machines. Our research focuses on enabling machines to both create and comprehend these complex language forms.
• Real-World Simulations: Recognizing that benchmarks often fall short of real-world dynamics, we propose solutions for home environments and robotics, ensuring actions are safe, adaptive, and equitable for all users.
Join Us
We invite researchers, industry partners, and aspiring students to collaborate with us. Together, we can push the boundaries of what’s possible in artificial intelligence and natural language processing. Whether you’re looking to advance cutting-edge research or develop practical applications, the Language Media Laboratory is the place to make it happen.
Recent topics
ナラティブセラピー用の対話処理
日本語の言語モデルのバイアス分析
輸出管理の法文書処理
文学における倫理判断
人工ディベーターのための自動要約
モバイル端末用の顔文字推薦
ゲーム内のNPCの対話による発話者推定
文脈による危険度の変化推定
チャットボット用の発話生成
オノマトペデータベースの自動構築
言い換え発見用の低コストな学習アルゴリズム
隠喩的表現の自動検出
チャットボット評価におけるAvatarの影響
駄洒落の面白さの自動判断
医療ニュース記事の根拠となる論文の候補抽出
日本語学習者のための文法自動修正
アスペクトベースの感情分析の改善
画像情報を用いた単語埋め込み
動物とのコミュニケーション
政治的な立場・プロパガンダ発見
レシピ生成
QAにおける常識処理
LML in Media
“Robot Be Good” (American Scientific) [和訳]
サイエンスZERO「料理！洗濯!笑いまで! 人と暮らす未来のロボット」 [ポスター][他のテレビ番組]
北海道新聞 「常識的知識や倫理AIで判断」
Stanford Rewired: Linguistic Revival: How Japan Restored the Native Ainu Language with “AI Pirika”
Hokkaido University Press: “The Vacuum Cleaner that Can Tweet It Can’t Fly”