openalex_id:w2963879515

著者: Lea Frermann, Shay B. Cohen, Mirella Lapata, Stanislaw Antol, Aishwarya Agrawal, Jiasen Lu, Margaret Mitchell, Dhruv Batra, C Zitnick, Devi Parikh, P Bojanowski, F Bach, I Laptev, J Ponce, C Schmid, J Sivic, John Boreczky, Lynn Wilcox, R Samuel, Gabor Bowman, Christopher Angeli, Christopher Potts, Manning, Elia Bruni, Nam Khanh Tran, Marco Baroni, Sachin Chachada, C.-C. Jay Kuo, Jacob Cohen, Timothee Cour, Chris Jordan, Eleni Miltsakaki, Ben Taskar, Steven Davis, Paul Mermelstein, Nevenka Dimitrova, Lalitha Agnihotri, Gang Wei, Desmond Elliott, Frank Keller, Philip John, Gorinski, Mirella Lapata, Karl Moritz Hermann, Tomas Kocisky, Edward Grefenstette, Lasse Espeholt, Will Kay, Mustafa Suleyman, Phil Blunsom, Curran Associates, Inc, Felix Hill, Anoine Bordes, Sumit Chopra, Jason Weston, Sepp Hochreiter, Jrgen Schmidhuber, Justin Johnson, Ranjay Krishna, Michael Stark, Li-Jia Li, David Shamma, Michael Bernstein, Li Fei-Fei, Andrej Karpathy, Li Fei-Fei, Douwe Kiela, Lon Bottou, John Lafferty, Andrew Mccallum, Fernando Pereira, Angeliki Lazaridou, Nghia The Pham, Marco Baroni, Dahua Lin, Sanja Fidler, Chen Kong, Raquel Urtasun, C Myers, L Rabiner, H Naphide, T Huang, Luis Gilberto, Mateos Ortiz, Clemens Wolff, Mirella Lapata, Jeffrey Pennington, Richard Socher, Christopher Manning, Pranav Rajpurkar, Jian Zhang, Konstantin Lopyrev, Percy Liang, Z Rasheed, Y Sheikh, M Shah, Matthew Richardson, J Christopher, Erin Burges, Renshaw, Tim Rocktschel, Edward Grefenstette, Karl Hermann, Tomas Kocisky, Phil Blunsom, Anna Rohrbach, Atousa Torabi, Marcus Rohrbach, Niket Tandon, Christopher Pal, Hugo Larochelle, Aaron Courville, Bernt Schiele, Md, Goutam Sahidullah, Saha, Jitao Sang, Changsheng Xu, Carina Silberer, Vittorio Ferrari, Mirella Lapata, Josef Sivic, Mark Everingham, Andrew Zisserman, Ilya Sutskever, Oriol Vinyals, Quoc Le, Christian Szegedy, Sergey Ioffe, Vincent Vanhoucke, Makarand Tapaswi, Martin Buml, Rainer Stiefelhagen, Makarand Tapaswi, Yukun Zhu, Rainer Stiefelhagen, Antonio Torralba, Raquel Urtasun, Sanja Fidler, Subhashini Venugopalan, Marcus Rohrbach, Jeff Donahue, Raymond Mooney, Trevor Darrell, Kate Saenko, Subhashini Venugopalan, Huijuan Xu, Jeff Donahue, Marcus Rohrbach, Raymond Mooney, Kate Saenko, Oriol Vinyals, Alexander Toshev, Samy Bengio, Dumitru Erhan, Ellen Voorhees, Dawn Tice, Jason Weston, Antoine Bordes, Sumit Chopra, Tomas Mikolov, Kelvin Xu, Jimmy Ba, Ryan Kiros, Kyunghyun Cho, Aaron Courville, Ruslan Salakhudinov, Rich Zemel, Yoshua Bengio, Yi Yang, Wen-Tau Yih, Christopher Meek, Mark Yatskar, Luke Zettlemoyer, Ali Farhadi, Wojciech Zaremba, Ilya Sutskever, Oriol Vinyals, Yukun Zhu, Ryan Kiros, Rich Zemel, Ruslan Salakhutdinov, Raquel Urtasun, Antonio Torralba, Sanja Fidler - University of Edinburgh 被引用: 25

In this paper we argue that crime drama exemplified in television programs such as CSI: Crime Scene Investigation is an ideal testbed for approximating real-world natural language understanding and the complex inferences associated with it.We propose to treat crime drama as a new inference task, capitalizing on the fact that each episode poses the same basic question (i.e., who committed the crime) and naturally provides the answer when the perpetrator is revealed.We develop a new dataset 1 based on CSI episodes, formalize perpetrator identification as a sequence labeling problem, and develop an LSTM-based model which learns from multi-modal data.Experimental results show that an incremental inference strategy is key to making accurate guesses as well as learning from representations fusing textual, visual, and acoustic input.