Digital Cultural Heritage Engagement

[nickumia]

• Safar: Heuristics for Augmented Reality Integration in Cultural Heritage (2023)
• Enhancing Engagement through Digital Cultural Heritage: A Case Study about Senior Citizens using a Virtual Reality Museum (2023)
• A Social Awareness Interface for Helping Immigrants Maintain Connections to Their Families and Cultural Roots: The Case of Venezuelan Immigrants (2023)
• Black Innovation: Creating and Appropriating Technology to Survive and Thrive (2023)
• SeMap: A project on semantisation and visualisation of cultural heritage data (2022)
• Poverty rate prediction using multi-modal survey and earth observation data (2023)
• “Do you collect data to give to the university or do you do the work to benefit people?”: Indigenous Data Sovereignty in Environmental Contexts (2023)
• Problematic smartphone usage, emotion regulation, and social and non-social smartphone use (2018)
• A Multi-factorial Analysis of Polarization on Social Media (2023)
• Factors Influencing the Willingness to Download Contact Tracing Apps among the American Population (2023)
• Computer Vision, Human Likeness, and Problematic Behaviors: Distinguishing Stereotypes from Social Norms (2023)
• A Day of Your Days: Estimating Individual Daily Journeys Using Mobile Data to Understand Urban Flow (2016)

Old Papers

• Using social networks for multicultural creative collaboration (2012)
• Knowing me knowing you: exploring effects of culture and context on perception of robot personality (2012)
• Detecting value differences behind intercultural meetings (2012)

[nickumia]

Safar: Heuristics for Augmented Reality Integration in Cultural Heritage (2023)

METHODOLOGY

• Research began with 20 participants.
• Questionnaire gathered baseline data about participants’ knowledge of cultural heritage sites + their experience with AR technologies
• Participants watched a virtual tour of the Qutub Minar, ensuring everyone had the same basic knowledge about the site
• Explored AR applications related to cultural heritage
• 12 participants was then interviewed to gather deeper insights about their thoughts on AR in cultural heritage contexts
  • Semi-structured interviews
• 9 participants were grouped into teams of three and guided through a Six-step research process
  • introduction to the central challenge
  • 15-minute "think aloud" session
  • organized on a quadrant graph, divided by effort on the x-axis (low to high) and impact on the y-axis (low to high)
  • participants sketched preliminary application screen designs on paper
  • reflective exercise, where participants responded to a set of prompts on their papers,

Analysis

• Cultural heritage serves as a link to individuals’ origins and roots
  • Extends beyond physical locations
  • Encompass the understanding of traditions and practices’ origins
  • Shapes both personal and communal identities
• Role of guides in heritage sites
  • Ability of guides to craft narratives
  • Accuracy of the information
  • Language barriers with guides were noted (multilingual guides or digital tools to bridge this gap and provide more inclusive experiences)
• preferences for self-exploration
  • lack of informative resources, like detailed maps or descriptions, was seen as a hindrance
• Incorporating additional attractions or activities in the vicinity of monuments
• 3D hologram guides that could add an interactive and personal dimension to exploration
• seating areas and shade
• visualize historical events
• explore reconstructions of damaged structures
• holographic guides for a more dynamic learning
• Personal curiosity and connections
• practicality and convenience were raised as considerations
  • lack of nearby amenities and attractions

Guidelines

• Ensure well-trained guides for accurate historical narratives.
• Verify information to prevent misrepresentation.
• Encourage emotionally resonant storytelling.
• Provide multilingual guides and digital aids.
• Offer translated materials and interactive apps.
• Balance guided tours with self-exploration.
• Incorporate Augmented Reality for personalized experiences.
• Tailor experiences to individual connections.
• Introduce attractions and cultural activities.
• Embrace a holistic perspective for storytelling and comfort.
• Create educational opportunities and hands-on experiences.
• Establish a feedback mechanism for continuous improvement.
• Implement sustainable practices and educate on conservation.
• Foster collaborations for a well-rounded experience.
• Utilize social media for virtual engagement.

[nickumia]

Computer Vision, Human Likeness, and Problematic Behaviors: Distinguishing Stereotypes from Social Norms (2023)

• Clarifai
• The issue is that these machines lack lived experience and thus, they have not been taught social norms
• Humans know – based on their social experiences – what is appropriate and meaningful to say about other people, and in which situations; this is missing from CV technologies.
• Machine Social Behaviors vs Problematic Machine Social Behaviors vs Social Stereotypes vs Social Norms vs Social Norms for Language
• For instance, in a dating app, the automated labeling of a person's image with a tag such as “handsome" or “elegant" may be viewed as harmless, perhaps even in a positive light.
• in the context of a human resources application (e.g., screening of candidates’ LinkedIn photos) the use of the inferential and/or demographic attributes, will likely be off limits to many observers.

Human Process	Related Clarifai Concepts
1. Describe elements composing the image	computer, coffee, man, woman
2. Analyze the arrangement of those elements	sitting, looking, indoors, college, room
3. Interpret the overall message the image conveys	togetherness, teamwork, family, cooperation
4. Appreciate the aesthetics of the image	N/A

• ethical debt of modern AI
• Three key shortcuts are described:
  • i) the emphasis on big data and machine learning models, which are based on correlation rather than causation;
  • ii) the use of data captured in the wild rather than bespoke training datasets;
  • iii) the use of proxies and implicit feedback to infer – rather than deeply understand – what users want
• we have taught cognitive technologies such as CV to mimic us, but we have not taught them social norms, in part because we do not all share the same social values and norms.
  • So, ultimately, we are the source of the biases that surface when algorithms are trained on our big data
• if our goal is human likeness, we must admit that social stereotyping is a reflection of this engineering goal, and that it must be managed rather than eradicated.
• Those of us who work in HCI should find creative ways of using machine behavior approaches to raise awareness in / educate users at large. As mentioned, having evidence of machine behaviors might better help users to calibrate their expectations as to what AI can and cannot do.

[nickumia]

Factors Influencing the Willingness to Download Contact Tracing Apps among the American Population (2023)

• perceived trust and perceived enjoyment are significant motivators (Canadian)

• perceived trust in app providers POSITIVE (german)

• perceived severity of data misuse and perceived vulnerability to data misuse NEGATIVE (german)

• financial and non-financial reward (e.g., permission to socialize), user-controlled data sharing and privacy protection POSITIVE (dutch)

• privacy concern, data security concern, non-ownership of phone, and out-of-pocket cost NEGATIVE (dutch)

• erceived usefulness, subjective norm, and innovativeness POSITIVE (belgians)

• privacy concern NEGATIVE (belgians)

• Germany and the United States were less supportive than the other countries

• rivacy concern, cybersecurity concern, and lack of trust in government NEGATIVE (France, Germany, Italy, United Kingdom, and the United States)

• What factors determine the willingness to download a CTA from the app stores among the American population?

• Do persuasive design and smartphone usage experience moderate the determinants of the willingness to download a CTA from the app stores by the American population?

• University of Waterloo's ethics office for review

• posted on Amazon Mechanical Turk (AMT) to recruit participants resident in America between January and February 2021.

• 242 participants resident in the United States

• randomly assigned one of the six test conditions

• US$ 2 in appreciation of their time

• excluding current CTA adopters

• participants who did not respond to the question on willingness to download

• 160 participants

• unified theory of acceptance and use of technology (UTAUT)

• Partial Least Square Path Modeling (PLSPM)

• However, the second significant construct (perceived compatibility) turns out to be a demotivator. This finding is counterintuitive, as one would think that the more the app is compatible with prior mobile apps used by the users, the more likely they are to adopt the app. Compared with the persuasive design model, in which perceived compatibility is non-significant, the finding indicates that the simpler (i.e., more minimalistic) an ENA (control design) is among the American population, the less likely they are to adopt it to curb the spread of the virus

• In general, the usefulness of the app (the utility it offers users) should be emphasized in marketing campaigns and app descriptions, e.g., in the app stores.

• For users with more years of experience in smartphone usage, the utility of the app as well as the confidentiality and privacy of users’ data should be emphasized. Moreover, for low-experience users, trust in the app should be fostered by describing how it protects the user's interest including their data.

• For apps that incorporate persuasive features such as self-monitoring that tracks exposure statistics such as number of contacts, sponsors should foster trust by describing the privacy measures to protect the user's data.

• The results of our path modeling show that perceived usefulness is the most and only important determinant of the target population's willingness to download a CTA from the app stores. This finding is the same for the group that evaluated the control designs. However, among the group that evaluated the persuasive designs, perceived trust is the most and only important factor that influences the willingness to download the app. Moreover, for users with more years of experience in smartphone usage, perceived usefulness and perceived risk are the important factors, while for users with less years of experience, perceived trust is the important factor.

[nickumia]

A Multi-factorial Analysis of Polarization on Social Media (2023)

Research questions

• RQ1: Do the current individual polarization metrics contribute to distinguishing polarized users from non-polarized users?
• RQ2: Can a multi-factorial analysis identify different classes of polarization behavior

Scores

• The polarization score of a user u depends on the ratio of interactions in each community, with Nu, c the number of interactions of u in community c, and Nu her total number of interactions
• Lack of Diversity (LD) metric; considers sources of information a user interacted with, concretely a set of M media outlets, rather than communities

Method

• Twitter API (v2), with academic research access
• elite users -- legitimacy: (1) have a significant number of followers; (2) personally manage their Twitter account; (3) are known by the general audience, through media or government interventions; (4) are qualified by education and/or profession to address the subject of matter.
• (1) Collect all tweets published by the set of elite users during the predefined period;
• (2) Filter tweets about the topic of interest;
• (3) Collect information about a random subset of interacting standard users for each collected tweet;
• (4) Identify the most active standard users among those selected in Step 3;
• (5) Collect all interactions of selected standard users on collected elite users’ tweets during the defined period.
• 20 French-speaking elite users having a legitimate voice in the vaccine debate (10 pro-vaccine and 10 anti-vaccine
• tweets between January 1, 2022 and July 31, 2022
• retweets, which are signs of approval and thus give information about what users agree with
• Among the selected retweeters, we focused on the 1,000 most active ones (500 pro-vaccine and 500 anti-vaccine).
• 6,697 tweets
  • 1,869 tweets from pro-vaccine elite
  • 4,828 tweets from anti-vaccine elite
• 1,000 most active retweeters,
  • 11,449,936 retweets
    • 299,879 of these retweets were on elite users’ tweets
      • 16,791 retweets on pro-vaccine tweets
      • 283,088 retweets on anti-vaccine tweets
• difficult to move from one community to the other one (Random Walk process equal to 0.89)
• Opinion factor, where opinions are assessed from the standard users’ retweets on each community
• Source factor, where sources are assessed from standard users’ retweets on each elite user (source)
• Single Factor Analysis, Kernel Density Estimation (KDE)
• Bi-factor Analysis, k-means
• Tri-factor Analysis,
• these results indicate that current polarization metrics do not distinguish polarization behaviors properly (RQ1), and that entropy-based metrics seem better adapted. Besides, conducting a tri-factor analysis allows an unprecedented identification of well-separated behavioral clusters, which emphasizes that an adequate combination of factors leads to more reliable modeling of polarization behaviors (RQ2).

[nickumia]

SeMap: A project on semantisation and visualisation of cultural heritage data (2022)

IT'S IN SPANISH 😖

[nickumia]

A Social Awareness Interface for Helping Immigrants Maintain Connections to Their Families and Cultural Roots: The Case of Venezuelan Immigrants (2023)

• Subtle and emotional communication, sometimes critical in close relationships, is underrepresented
• include the rituals and practices of the cultures that are using technology, rather than simply focusing on the technology
• focuses on Venezuelan immigrants.
• design a solution that mitigates loneliness by enhancing remote communication when the user experiences negative feelings due to separation
• KEPEIN, a coffee mug-like communication interface designed to enhance social awareness and connectivity by recreating the coffee drinking experience in remote communication contexts
• InTouch [12], enables social relationships for seniors separated from their families
• Whisper Pillow [7] is an interactive artifact for mediating emotional expression among couples with different daily routine
• Snowglobe [23] enhanced salience and closeness
• Tangiball [14] led to an enhanced social experience and a stronger sense of presence
• Huggy Pajama /cite[conect3] allows remote physical interaction through two physical entities connected via the Internet
• Messaging Kettle [4] encourages communication with faraway friends

• most salient keywords were: connectivity, privacy, sharing, touch, smell, memories, routines

[nickumia]

The culmination of this exploration for my assignment:

• COT6930_Kumia_HW4.pdf
• COT6930_Kumia_HW4_Presentation.pdf

[nickumia]

Youtube Video of Presentation: https://youtu.be/r_8S9chJ6bw?si=Iir56_cKYypb60tS