【正文】 speaking turn: stop vocalizing and reestablish eye contact.? To hold the floor: look to the side.? To relinquish the floor: raise brows and/or lean forward a bit.“使命”的信封，顯示如下：? 獲得說話的權(quán)利：打破目光接觸和/或稍微向后傾斜? 一開始的發(fā)言：發(fā)出者的含糊不清的話。 《計算機專業(yè)英語》課程論文13? 停止說話：停止發(fā)聲和重新建立眼神的接觸。? 演講：看著一邊。? 結(jié)束演講：提高眉毛和/或稍微向前傾斜。Blinking tends to occur at the end of a vocalization.閃爍往往一個發(fā)聲結(jié)束后出現(xiàn)。. Vocal turntaking experiments 聲音開啟試驗To investigate Kismet’s turntaking performance during protodialogs, we invited four naive subjects to interact with Kismet. Subjects ranged in age from 12 to 28 years old. Two male and two female subjects participated. In each case, the subject was simply asked to carry on a “play” conversation with the robot. The exchanges were video recorded for later analysis and annotated according to Table 1. The subjects were told that the robot neither speaks nor understands English, but babbles in a characteristic manner. Due to space limitations here, only a small portion of the protodialog carried out by Subject C (a female subject) is presented in Appendix A. The time codes are those that appear on the videotape used to record the sessions. A turn is defined with respect to the speaker who holds the floor and consists of four phases: acquire the floor (Aq), start the utterance (St), stop the utterance (Sp), and relinquish the floor (Rq). The speaker may also hold the floor (Hd), ., maintain their speaking role during a silent pause.在原對話框里面探討“使命”的話輪轉(zhuǎn)換表現(xiàn)，我們邀請了四個自然主體與“使命”互動。受試者的年齡從 12 歲到 28 歲。兩個男的和兩個女的受試者參加。在每一種情況下，主體只是要求進行與機器人 “玩”的對話。根據(jù)表1，為了之后的分析和注釋，交流的內(nèi)容被錄像機錄制下來了。受試者被告知說機器人既不說話也不懂英語，但在一個獨特的方式下牙牙學(xué)語。由于空間的限制，只有主體 C（女性主體）進行的原對話的一小部分是出現(xiàn)在附錄 A 中。時間碼即那些用于記錄會話的錄像帶。一個回合就誰擁有地板揚聲器的定義包括四個階段：獲取地板（AQ） ，啟動話語（ST） ，停止話語（SP） ，并且放棄地板（RQ） 。發(fā)言人還可以把地板（HD） ，即，一種無聲停頓期間保持他們的說話的角色。Table 1. Annotations for protodialog experiments (see Appendix A for sample of annotated video)表 1。原對話框?qū)嶒炞⑨專ㄒ姼戒浺蛔⑨尩囊曨l樣本） 《計算機專業(yè)英語》課程論文14Type Option AnnotationsListener, speaker Human HRobot RTurn phase Acquire floor AqStart speech StStop speech SpHold floor HdRelinquish floor RqCue Avert gazeEye contactElevate browsLean forwardLean backBlink“Utterance”Turns Clean turn Interrupt IMissed MPause P. Evaluation 評價We evaluate these human–robot interactions with respect to three criteria. These are inherently subjective, yet quantifiable, measures that evaluate the quality and ease of interaction between human and robot. They address the behavior of both partners, human and robot. The evaluation criteria are as follows:我們評估這些人與機器人之間的相互作用方面的三個標準。這些都是主觀的和可以計量的對人類與機器人之間的相互作用的測量。他們專注于人類與機器人雙方的行為。評價標準如下： 《計算機專業(yè)英語》課程論文15? Do people intuitively read and naturally respond to Kismet’s social cues?? Can Kismet perceive and appropriately respond to these naturally offered cues?? Does the human adapt to the robot, and the robot adapt to the human, in a way that benefits the interaction? Specifically, is the resulting interaction natural, intuitive, and enjoyable for the human。 can Kismet perform well despite its perceptual, mechanical, behavioral, and putational limitations?? 人們能直觀讀取和自然回應(yīng)命運的社會線索嗎？?“使命”能感知和妥善應(yīng)對這些自然提供的線索嗎？? 是人類適應(yīng)機器人，和機器人適應(yīng)人類，在某種程度上有利于相互作用？具體 地說，相互作用產(chǎn)生的使人自然的，直觀的，和愉快的；除了機械，行為，和 計算的限制， “使命”能很好地表現(xiàn)嗎？Often the subjects begin the session by speaking longer phrases and only using the robot’s vocal behavior to gauge their speaking turn. They also expect the robot to respond immediately after they finish talking. Before the subjects adapt their behavior to the robot’s capabilities, the robot is more likely to interrupt them. For instance, it is often the case that the robot interrupts them within the first couple of exchanges. In general, there tends to be more frequent delays in the flow of “conversation” where the human prompts the robot again for a response. Often these “hiccups” in the flow appear in short clusters of mutual interruptions and pauses (often over 2–4 utterances of the speaker) before the turn phases bee coordinated and the flow of the exchange of speaking turns smoothes out. We call these clusters significant flow disturbances.受試者一開始經(jīng)常會說較長的詞組并且只使用機器人的發(fā)聲行為評估他們的發(fā)言。他們也希望機器人在他們說完后立即回應(yīng)。在受試者使得他們的行為適應(yīng)機器人的能力之前，機器人更容易打斷他們。例如，機器人通常在第一段對話中打斷他們。在一般情況下，人類再次提示機器人回應(yīng)的 “對話”更加頻繁的延遲響應(yīng)。通常這些“打嗝” 出現(xiàn)在短簇相互中斷和停頓（通常超過 2–4 話語的說話人） ，轉(zhuǎn)前階段成為協(xié)調(diào)和講輪流交換流量平滑。我們稱為這些集群的顯著的流動擾動。Video analysis of these human–robot “conversations”, provide evidence that 《計算機專業(yè)英語》課程論文16people entrain to the robot (seeTable 2). They often start to use shorter phrases, wait longer for the robot to respond, and more carefully watch the robot’s turntaking cues. For instance, the robot prompts the person to take their speaking turn by either craning its neck forward, raising its brows, or establishing eye contact when it is ready for them to speak. It will hold this posture for a few seconds until the person responds. Often, within a second of this display, the subject does so. When the subject stops speaking, Kismet tends to lean back to a neutral posture, assume a neutral expression, and perhaps shift its gaze away from the person. This cue indicates that the robot is about to speak. The robot typically issues one utterance, but it may issue several. Noheless, as the exchange proceeds, the subjects are more likely to wait until prompted by the relinquish turn display.這些人–機器人“對話”的視頻分析提供的證據(jù)表明，人喜歡機器人（見表 2） 。他們開始使用較短的詞組，花更多的時間等待機器人的回應(yīng)和更仔細地觀察機器人的轉(zhuǎn)向線索。例如，當(dāng)它們準備好說話，機器人會提示人們輪流講話時伸長脖子向前，抬起眉毛，或建立目光接觸。它將保持這個姿勢幾秒鐘直到人們作出反應(yīng)。通常，一秒鐘內(nèi)機會顯示主題的內(nèi)容。當(dāng)主體停止說話， “使命傾向于回到中立姿態(tài)，假設(shè)一個中立的表達，并可能改變其目光離開的人。這個線索表明機器人想說話。機器人話語只是問題之一，它也有可能其他更多的問題。盡管如此，隨著交換收益，受試者更可能等到被提示要結(jié)束對話。Table 2. Data illustrating evidence for entrainment of human to robot (as time progresses there are increasing number