Showing 1–11 of 11 results for author: Kurihara, K

LCTG Bench: LLM Controlled Text Generation Benchmark

Authors: Kentaro Kurihara, Masato Mita, Peinan Zhang, Shota Sasaki, Ryosuke Ishigami, Naoaki Okazaki

Abstract: The rise of large language models (LLMs) has led to more diverse and higher-quality machine-generated text. However, their high expressive power makes it difficult to control outputs based on specific business instructions. In response, benchmarks focusing on the controllability of LLMs have been developed, but several issues remain: (1) They primarily cover major languages like English and Chines… ▽ More The rise of large language models (LLMs) has led to more diverse and higher-quality machine-generated text. However, their high expressive power makes it difficult to control outputs based on specific business instructions. In response, benchmarks focusing on the controllability of LLMs have been developed, but several issues remain: (1) They primarily cover major languages like English and Chinese, neglecting low-resource languages like Japanese; (2) Current benchmarks employ task-specific evaluation metrics, lacking a unified framework for selecting models based on controllability across different use cases. To address these challenges, this research introduces LCTG Bench, the first Japanese benchmark for evaluating the controllability of LLMs. LCTG Bench provides a unified framework for assessing control performance, enabling users to select the most suitable model for their use cases based on controllability. By evaluating nine diverse Japanese-specific and multilingual LLMs like GPT-4, we highlight the current state and challenges of controllability in Japanese LLMs and reveal the significant gap between multilingual models and Japanese-specific models. △ Less

Submitted 27 January, 2025; originally announced January 2025.

Comments: 15 pages, 11 figures. Project page: this [URL](https://github.com/CyberAgentAILab/LCTG-Bench)

GameControllerizer: Middleware to Program Inputs for Augmenting Digital Games

Authors: Kazutaka Kurihara, Nobuhiro Doi

Abstract: This study proposes middleware, GameControllerizer, that allows users to combine the processes of Internet of Things (IoT) devices, Web services, and applications of Artificial Intelligence (AI), and to convert them into game control operations to augment existing digital games. The system facilitates easy trial-and-error development of new forms of entertainment and the configuration of gamificat… ▽ More This study proposes middleware, GameControllerizer, that allows users to combine the processes of Internet of Things (IoT) devices, Web services, and applications of Artificial Intelligence (AI), and to convert them into game control operations to augment existing digital games. The system facilitates easy trial-and-error development of new forms of entertainment and the configuration of gamification by enabling the use of diverse devices and sources of information as inputs to games. GameControllerizer consists of a visual programming element that uses the Node-RED tool to allow users to program easily to convert diverse formats of information into inputs to games, and contains a game input emulation element whereby hardware- and software-based emulation generates inputs for gaming devices. Evidence of the usefulness of the system was provided by a performance assessment and the proposal of a variety of use cases. △ Less

Submitted 2 October, 2018; originally announced October 2018.

AnimalCatcher: a digital camera to capture various reactions of animals

Authors: Koji Tsukada, Maho Oki, Kazutaka Kurihara, Yuko Furudate

Abstract: People often have difficulty to take pictures of animals, since animals usually do not react with cameras nor understand verbal directions. To solve this problem, we developed a new interaction technique, AnimalCatcher, which can attract animals' attention easily. The AnimalCatcher shoots various sounds using directional speaker to capture various reactions of animals. This paper describes concept… ▽ More People often have difficulty to take pictures of animals, since animals usually do not react with cameras nor understand verbal directions. To solve this problem, we developed a new interaction technique, AnimalCatcher, which can attract animals' attention easily. The AnimalCatcher shoots various sounds using directional speaker to capture various reactions of animals. This paper describes concepts, implementation, and example pictures taken in a zoo. △ Less

Submitted 24 April, 2015; originally announced August 2015.

Comments: Written in Japanese, 6 pages

Quantum Annealing for Variational Bayes Inference

Authors: Issei Sato, Kenichi Kurihara, Shu Tanaka, Hiroshi Nakagawa, Seiji Miyashita

Abstract: This paper presents studies on a deterministic annealing algorithm based on quantum annealing for variational Bayes (QAVB) inference, which can be seen as an extension of the simulated annealing for variational Bayes (SAVB) inference. QAVB is as easy as SAVB to implement. Experiments revealed QAVB finds a better local optimum than SAVB in terms of the variational free energy in latent Dirichlet al… ▽ More This paper presents studies on a deterministic annealing algorithm based on quantum annealing for variational Bayes (QAVB) inference, which can be seen as an extension of the simulated annealing for variational Bayes (SAVB) inference. QAVB is as easy as SAVB to implement. Experiments revealed QAVB finds a better local optimum than SAVB in terms of the variational free energy in latent Dirichlet allocation (LDA). △ Less

Submitted 9 August, 2014; originally announced August 2014.

Comments: Appears in Proceedings of the Twenty-Fifth Conference on Uncertainty in Artificial Intelligence (UAI2009)

Report number: UAI-P-2009-PG-479-486

Quantum Annealing for Clustering

Authors: Kenichi Kurihara, Shu Tanaka, Seiji Miyashita

Abstract: This paper studies quantum annealing (QA) for clustering, which can be seen as an extension of simulated annealing (SA). We derive a QA algorithm for clustering and propose an annealing schedule, which is crucial in practice. Experiments show the proposed QA algorithm finds better clustering assignments than SA. Furthermore, QA is as easy as SA to implement. This paper studies quantum annealing (QA) for clustering, which can be seen as an extension of simulated annealing (SA). We derive a QA algorithm for clustering and propose an annealing schedule, which is crucial in practice. Experiments show the proposed QA algorithm finds better clustering assignments than SA. Furthermore, QA is as easy as SA to implement. △ Less

Submitted 9 August, 2014; originally announced August 2014.

Comments: Appears in Proceedings of the Twenty-Fifth Conference on Uncertainty in Artificial Intelligence (UAI2009)

Report number: UAI-P-2009-PG-321-328

A Face-like Structure Detection on Planet and Satellite Surfaces using Image Processing

Authors: Kazutaka Kurihara, Masakazu Takasu, Kazuhiro Sasao, Hal Seki, Takayuki Narabu, Mitsuo Yamamoto, Satoshi Iida, Hiroyuki Yamamoto

Abstract: This paper demonstrates that face-like structures are everywhere, and can be de-tected automatically even with computers. Huge amount of satellite images of the Earth, the Moon, the Mars are explored and many interesting face-like structure are detected. Throughout this fact, we believe that science and technologies can alert people not to easily become an occultist. This paper demonstrates that face-like structures are everywhere, and can be de-tected automatically even with computers. Huge amount of satellite images of the Earth, the Moon, the Mars are explored and many interesting face-like structure are detected. Throughout this fact, we believe that science and technologies can alert people not to easily become an occultist. △ Less

Submitted 13 June, 2013; originally announced June 2013.

Comments: 4 pages

Journal ref: ACE 2013, LNCS 8253, Springer, pp. 564-567, 2013

SpeechJammer: A System Utilizing Artificial Speech Disturbance with Delayed Auditory Feedback

Authors: Kazutaka Kurihara, Koji Tsukada

Abstract: In this paper we report on a system, "SpeechJammer", which can be used to disturb people's speech. In general, human speech is jammed by giving back to the speakers their own utterances at a delay of a few hundred milliseconds. This effect can disturb people without any physical discomfort, and disappears immediately by stop speaking. Furthermore, this effect does not involve anyone but the speake… ▽ More In this paper we report on a system, "SpeechJammer", which can be used to disturb people's speech. In general, human speech is jammed by giving back to the speakers their own utterances at a delay of a few hundred milliseconds. This effect can disturb people without any physical discomfort, and disappears immediately by stop speaking. Furthermore, this effect does not involve anyone but the speaker. We utilize this phenomenon and implemented two prototype versions by combining a direction-sensitive microphone and a direction-sensitive speaker, enabling the speech of a specific person to be disturbed. We discuss practical application scenarios of the system, such as facilitating and controlling discussions. Finally, we argue what system parameters should be examined in detail in future formal studies based on the lessons learned from our preliminary study. △ Less

Submitted 13 April, 2012; v1 submitted 27 February, 2012; originally announced February 2012.

ACM Class: H.5.m

CinemaGazer: a System for Watching Video at Very High Speed

Authors: Kazutaka Kurihara

Abstract: This paper presents a technology that enables the watching of videos at very high speed. Subtitles are widely used in DVD movies, and provide useful supplemental information for understanding video contents. We propose a "two-level fast-forwarding" scheme for videos with subtitles, which controls the speed of playback depending on the context: very fast during segments without language, such as su… ▽ More This paper presents a technology that enables the watching of videos at very high speed. Subtitles are widely used in DVD movies, and provide useful supplemental information for understanding video contents. We propose a "two-level fast-forwarding" scheme for videos with subtitles, which controls the speed of playback depending on the context: very fast during segments without language, such as subtitles or speech, and "understandably fast" during segments with such language. This makes it possible to watch videos at a higher speed than usual while preserving the entertainment values of the contents. We also propose "centering" and "fading" features for the display of subtitles to reduce fatigue when watching high-speed video. We implement a versatile video encoder that enables movie viewing with two-level fast-forwarding on any mobile device by specifying the speed of playback, the reading rate, or the overall viewing time. The effectiveness of our proposed method was demonstrated in an evaluation study. △ Less

Submitted 10 April, 2012; v1 submitted 4 October, 2011; originally announced October 2011.

Comments: 8 pages, 9 figures

ACM Class: H.5.1; H.5.2

Localizing Audiences' Gaze using a Multi-touch Electronic Whiteboard with sPieMenu

Authors: Kazutaka Kurihara, Naoshi Nagano, Yuta Watanabe, Yuichi Fujimura, Akinori Minaduki, Hidehiko Hayashi, Yohei Tsuchiya

Abstract: Direct-touch presentation devices such as touch-sensitive electronic whiteboards have two serious problems. First, the presenter's hand movements tend to distract the audience's attention from content. Second, the presenter' s manipulation tends to obscure content. In this paper we describe a new electronic whiteboard system that supports multi-touch gestures and employs a special pie menu interfa… ▽ More Direct-touch presentation devices such as touch-sensitive electronic whiteboards have two serious problems. First, the presenter's hand movements tend to distract the audience's attention from content. Second, the presenter' s manipulation tends to obscure content. In this paper we describe a new electronic whiteboard system that supports multi-touch gestures and employs a special pie menu interface named "sPieMenu." This pie menu is displayed under the presenter's palm and is thus invisible to the audience. A series of experiments shows that the proposed system allows both novice and expert users to efficiently manipulate the electronic whiteboard, and that the proposed system decreases distraction to the audience compared to traditional approaches. △ Less

Submitted 7 December, 2010; originally announced December 2010.

ACM Class: H.5.2

Quantum Annealing for Variational Bayes Inference

Authors: Issei Sato, Kenichi Kurihara, Shu Tanaka, Hiroshi Nakagawa, Seiji Miyashita

Abstract: This paper presents studies on a deterministic annealing algorithm based on quantum annealing for variational Bayes (QAVB) inference, which can be seen as an extension of the simulated annealing for variational Bayes (SAVB) inference. QAVB is as easy as SAVB to implement. Experiments revealed QAVB finds a better local optimum than SAVB in terms of the variational free energy in latent Dirichlet… ▽ More This paper presents studies on a deterministic annealing algorithm based on quantum annealing for variational Bayes (QAVB) inference, which can be seen as an extension of the simulated annealing for variational Bayes (SAVB) inference. QAVB is as easy as SAVB to implement. Experiments revealed QAVB finds a better local optimum than SAVB in terms of the variational free energy in latent Dirichlet allocation (LDA). △ Less

Submitted 28 May, 2009; v1 submitted 21 May, 2009; originally announced May 2009.

Comments: 9 pages, 4 figures, Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI 2009) accepted

Quantum Annealing for Clustering

Authors: Kenichi Kurihara, Shu Tanaka, Seiji Miyashita

Abstract: This paper studies quantum annealing (QA) for clustering, which can be seen as an extension of simulated annealing (SA). We derive a QA algorithm for clustering and propose an annealing schedule, which is crucial in practice. Experiments show the proposed QA algorithm finds better clustering assignments than SA. Furthermore, QA is as easy as SA to implement. This paper studies quantum annealing (QA) for clustering, which can be seen as an extension of simulated annealing (SA). We derive a QA algorithm for clustering and propose an annealing schedule, which is crucial in practice. Experiments show the proposed QA algorithm finds better clustering assignments than SA. Furthermore, QA is as easy as SA to implement. △ Less

Submitted 28 May, 2009; v1 submitted 21 May, 2009; originally announced May 2009.

Comments: 8 pages, 6 figures, Proceedings of the 25th Conference on Uncertainty in Artificial Intelligence (UAI 2009) accepted