Datasets

Date: September 18, 2024
Nº of Users: 215

Description: The dataset contains the results of a questionnaire conducted with 215 participants, who watched a series of videos of an elderly person interacting with the social robot SHARA in various scenarios. Participants were asked based on their perceptions of different levels of the robot’s proactivity (Presence, Dialogue, Suggestions, and Autonomy) presented through video simulations. The questionnaire included demographic questions (age, gender, familiarity with technology, etc.), as well as ratings on the usefulness, appropriateness, intrusiveness, and naturalness of the robot’s interactions for each scenario (Likert scale from 1 to 5, being 3 neutral).

In addition to the main questions, participants were also asked about privacy concerns, their comfort with the robot’s behavior, and whether they found the robot’s actions appropriate for elderly care. These questions were rated on a scale from 1 to 5, with higher values indicating greater concern or acceptance. For Dialogue, Suggestions and Autonomy proactivity levels, participants were also asked for which Instrumental Activities of Daily Living it was appropriate for the robot to have this level of proactivity (Likert from 1 to 5, being 3 neutral).

Finally, participants made an overall assessment using the Short User Experience Questionnaire (UEQ-S). 

Questionnaire available at: https://forms.gle/VJyymiLQQenmG9K77

Date: March 30, 2024
Description: it consists of two datasets, used to train models for intent classification and entity recognition. They have been created tailored to the input format for OpenAI queries: each entry (called 'message') is a dictionary in the conversational chat required by GPT-3.5-turbo models. This dictionary format has three elements, corresponding to the prompt (instructions on what the model should do and the output format), sample user input (in spanish), and model output.
(1) Dataset for intent classification. It has examples of user input and corresponding intent (can be ['Yes', 'No', 'Thanks', 'Make_Appointment', 'Know_Operation_Hours', 'Know_Location', 'Help', 'Greetings', 'Goodbye', 'Connect_to_agent', 'Irrelevant']). Contains 1060 entries (80% for training, with approximately 100 examples of each intent; and 20% test). (2) Dataset for entity extraction. It has examples of user entries and their extracted entities (of type time 'HH:mm', date 'yyyy-Mon-dd', and holiday) in dictionary form. It contains 300 entries (80% for train, with 250 entries; and 20% test). The prompt of each entry includes the reference timestamp to know the time of the example and to be able to extract calendar-dependent date and time entities.

Date: December 30, 2022
Nº of Users: 30
Description: The dataset has a population of n = 30 (33.33% female and 66.67% male), without any gait pathology and with a mean age of 44.27 ± 19.55. The subjects had the following anthropometric measurements taken prior to testing: height, foot length, shoulder height, elbow span, wrist extension, knee height and ankle height. The experiment was carried out in a 24 m long and 3 m wide corridor with a wireless human motion tracker at the School of Computer Science of the University of Castilla-La Mancha, Ciudad Real. The system developed by Xsens, consists of 15 devices called (Motion Tracker Wireless), which can synchronize with a transmitter/receiver base (Awinda Station). The Awinda protocol uses 2.4 GHz and is based on the IEEE 802.15.4, with an accuracy of 10 us at a frequency of 60 Hz using 15 devices. Participants performed 15 activities (1 single task, 2 traditional dual-tasks and 12 mobile-based dual-tasks) while walking.

Date: October 17, 2018
Nº of Users: 10
Description: The dataset is published in a one CSV per participant with 24 rows (one per dyad conversation, plus the header) and 10 columns for the different variables. The description of the 10 columns is as follows. The first column contains the iteration number that corresponds to the rest of the data. The second has the subject of the question, as per the emotion perception tryad (one’s emotional self-awareness, empathy and social-emotional skills), and the third shows if it was a direct or indirect question. The fourth column shows the expected emotion that is returned by the user’s answer, while the fifth shows the actual emotion detected. The sixth column contains the transcription by Watson’s Speech-To-Text of what the user said, the seventh column contains the translation to English of that transcription, and the eighth the question that was asked by the system. Finally, the ninth column is blank, to be filled with expert observations, and the tenth has the latency for each iteration. The dataset does not contain any identifying aspects of the users

Date: October 9, 2019
Nº of Users: 6
Description: Participants performed 12 tasks considered in the HuSBIT-10 taxonomy. EEG activity was recorded during each tasks for a 10-second interval (EEG segment). The task performed are: C1 Read a message that contains a poem, C2 Listen to a podcast from the daily news, C3 Watch a video, E1 Search for a given date in the calendar, E3 Switch off the data roaming in the device settings, E4 Search how to reach a given place (about 500 meters away) in the map from the current location, M2 Add and move an app shortcut (2 times), M4 Copy a message into the browser search box (Google widget), M5 Select one word, then two and, finally, two and a half words in a Wikipedia article, P1 Write down the places where you would go in a zombie apocalypse, P2 Create a voice message with the list of objects you would collect in a zombie apocalypse, P3 Take an artistic photo of one object around you.

Date: February 7, 2020
Nº of Users: 26
Description: Participants performed 12 tasks considered in the HuSBIT-10 taxonomy. EEG activity was recorded during each tasks for a 10-second interval (EEG segment). The task performed are: C1 Read a message, C2 Listen to a voice recording, C3 Watch a video, E2 Count the number of beach umbrellas in a picture from Where's Wally?, E3 Switch off the data roaming in the device settings, E4 Search how to reach a given place (about 500 meters away) in the map from the current location, M2 Add and move an app shortcut (2 times), M4 Close all the Apps in the brackground, M5 Select one word, then two and, finally, two and a half words in a Wikipedia article, P1 Write down an excuse or justification for not attending with a meeting with someone, P2 Create a voice message answering a friend who has just written about cancelling the meeting, P3 Take an artistic photo of several object around you.

Date: August 13, 2021
Nº of Users: 19
Description: This dataset contains data gathered during an experiment in which 19 participants completed n-Back test for n 1,2 & 3 and Stroop test for congruent color and text and incongruent color and text. The data contains is conformed by a EEG signals, as well as events during the tests. Every data is contained in .csv files and there are three types of files: EEG signals, n-back test events and Stroop test events