Introduction Framework Experiments Conclusions Modeling Vocal Interaction for Text-Independent Participant Characterization in Multi-Party Conversation Kornel Laskowski 1 , 2 , Mari Ostendorf 3 & Tanja Schultz 1 , 2 1 Cognitive Systems Labs, Universit¨ at Karlsruhe 2 Language Technologies Institute, Carnegie Mellon University 3 Dept. Electrical Engineering, University of Washington June 20, 2008 K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Vocal Interaction (Dabbs & Ruback, 1987) vocal activity patterns for all K participants, seen together V ≡ { � , � } ≡ { 0 , 1 } V K talkspurt start/end times = text-independence at time t , vocal activity of participant k : q t [ k ] ∈ entire K -participant conversation: q t ∈ we’ll use a discretized version (frame step = 200 ms) K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Participant Characterization Jane a useful partition of the Mary conversation participants C ≡ h ( C ) role Joe influence seniority Fred dominance ranking (of the above) Sam for all time, the class of participant k : g [ k ] ∈ C ≡ { C 1 , · · · , C N } K -participant group: g ∈ K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Participant Characterization Jane a useful partition of the Mary conversation participants C ≡ h ( C ) role Joe influence seniority Fred dominance ranking (of the above) Sam for all time, the class of participant k : g [ k ] ∈ C ≡ { C 1 , · · · , C N } K -participant group: g ∈ K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Participant Characterization Jane a useful partition of the Mary conversation participants C ≡ h ( C ) role Joe influence seniority Fred dominance ranking (of the above) Sam for all time, the class of participant k : g [ k ] ∈ C ≡ { C 1 , · · · , C N } K -participant group: g ∈ K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Participant Characterization Jane a useful partition of the Mary conversation participants C ≡ h ( C ) role Joe influence seniority Fred dominance ranking (of the above) Sam for all time, the class of participant k : g [ k ] ∈ C ≡ { C 1 , · · · , C N } K -participant group: g ∈ K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions Participant Characterization Jane a useful partition of the Mary conversation participants C ≡ h ( C ) role Joe influence seniority Fred dominance ranking (of the above) Sam for all time, the class of participant k : g [ k ] ∈ C ≡ { C 1 , · · · , C N } K -participant group: g ∈ K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions What we’re trying to do 1 1 V K × T 2 3 F 3 4 2 4 { q t } ∈ g ∈ h ( C ) 1 given a sequence of T K -participant states q t 2 compute & model features F 3 infer required equivalence classes g [ k ] of each participant K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions What we’re trying to do 1 1 V K × T 2 3 F 3 4 2 4 { q t } ∈ g ∈ h ( C ) 1 given a sequence of T K -participant states q t 2 compute & model features F 3 infer required equivalence classes g [ k ] of each participant K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Introduction Framework Experiments Conclusions What we’re trying to do 1 1 V K × T 2 3 F 3 4 2 4 { q t } ∈ g ∈ h ( C ) 1 given a sequence of T K -participant states q t 2 compute & model features F 3 infer required equivalence classes g [ k ] of each participant K. Laskowski, M. Ostendorf, T. Schultz SIGdial 2008, Columbus OH, USA
Recommend
More recommend
