Proceedings Article10.1145/1330572.1330577
Learning object-manipulation verbs for human-robot communication
Komei Sugiura,Naoto Iwahashi +1 more
- 15 Nov 2007
- pp 32-38
TL;DR: A machine learning method for mapping object-manipulation verbs with sensory inputs and motor outputs that are grounded in the real world and combines HMMs to generate trajectories to accomplish goal-oriented tasks is proposed.
read more
Abstract: This paper proposes a machine learning method for mapping object-manipulation verbs with sensory inputs and motor outputs that are grounded in the real world. The method learns motion concepts demonstrated by a user and generates a sequence of motions, using reference-point-dependent probability models. Four components, needed to learn objectmanipulation verbs, are estimated from camera images; (1) a trajector and landmark, which are the objects of transitive verbs; (2) a reference point; (3) an intrinsic coordinate system; and (4) parameters of the motion's probabilistic model. The motion concepts are learned using hidden Markov models (HMMs). In the motion generation phase, our method then combines HMMs to generate trajectories to accomplish goal-oriented tasks. Results from simulation experiments in which our method generates motion by combining learned motion primitives are shown.
read more
Chat with Paper
AI Agents for this Paper
Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps
Citations
Learning, Generation and Recognition of Motions by Reference-Point-Dependent Probabilistic Models
TL;DR: This paper presents a novel method for learning object manipulation such as rotating an object or placing one object on another using reference-point-dependent probabilistic models, which can be used for the generation and recognition of motions.
40
Learning Action Primitives
Dana Kulic,Danica Kragic,Volker Krüger +2 more
- 01 Jan 2011
TL;DR: This chapter overviews the recent approaches for learning and modeling action primitives for human and robot action and describes common approaches such as stochastic methods and dynamical systems approaches.
25
Active learning of confidence measure function in robot language acquisition framework
Komei Sugiura,Naoto Iwahashi,Hideki Kashioka,Satoshi Nakamura +3 more
- 03 Dec 2010
TL;DR: A method for estimating ambiguity in the commands by introducing an active learning framework with Bayesian logistic regression to human-robot spoken dialogue and conducting physical experiments in which a user and a manipulator-based robot communicated in spoken language to manipulate toys is proposed.
Situated Spoken Dialogue with Robots Using Active Learning
TL;DR: A method for estimating ambiguity in commands is proposed by introducing an active learning scheme with Bayesian logistic regression to human–robot spoken dialogue and asking confirmation questions before the execution of any motion to decrease the risk of accident.
18
Hierarchical Implicit Feedback Structure in Passive Dynamic Walking
Yasuhiro Sugimoto,Koichi Osuka +1 more
TL;DR: This paper derives an analytical Poincare map for 2-period walking and discusses the stability of PDW with this map and points out that there is a similar interesting structure in this PoincARE map.
17
References
The human semantic potential: Spatial language and constrained connectionism.
TL;DR: Part 1 Introduction: matter and method space and semantic potential negative evidence and language learning the modelling challenge constrained connectionism a brief exchange an overview of the book.
469
Speech parameter generation from HMM using dynamic features
Keiichi Tokuda,Takao Kobayashi,Satoshi Imai +2 more
- 09 May 1995
TL;DR: It is shown that the parameter generation from HMMs using the dynamic features results in searching for the optimum state sequence and solving a set of linear equations for each possible state sequence.
324
Grounding words in perception and action: computational insights
TL;DR: An exciting implication for cognitive modeling is the use of grounded systems to 'step into the shoes' of humans by directly processing first-person-perspective sensory data, providing a new methodology for testing various hypotheses of situated communication and learning.
196