Dataset Viewer
Auto-converted to Parquet Duplicate
lecture
stringclasses
575 values
image_name
stringlengths
21
21
source
stringlengths
6
13.2k
target
stringlengths
10
8.62k
image
imagewidth (px)
480
1.28k
CHI-A48F7
CHI-A48F7-0011000.jpg
Hi, I'm Zara Astrov, and I'm here to present the work we did on effects of communication directionality in AI agent differences in
أهلاً، أنا زارا أستروف، وموجودة هنا عشان أقدم العمل اللي عملناه عن تأثيرات اتجاهية التواصل في فروقات وكلاء الذكاء الاصطناعي في ---
CHI-A48F7
CHI-A48F7-0050455.jpg
human-AI interaction. HCI researchers have done a lot of work on determining what can lead to successful collaborations in human-AI collaborative settings. One factor that has been explored are social perceptions between collaborators. One opportunity for studying user perceptions of AI agents are cooperative partially observable games. These games require players to cooperate with one another, especially since there are aspects of the game that are only partially observable, so here the AI. AI agent has a target word, cold, that the user does not know about. But the AI agent will send a clue, and the user must interpret that.
التفاعل بين الإنسان والذكاء الاصطناعي. باحثين التفاعل بين الإنسان والكمبيوتر (HCI) عملوا شغل كتير عشان يحددوا إيه اللي ممكن يؤدي لتعاون ناجح في بيئات التعاون بين البشر والذكاء الاصطناعي. واحد من العوامل اللي تم استكشافها هو التصورات الاجتماعية بين المتعاونين. واحدة من الفرص لدراسة تصورات المستخدمين لوكلاء الذكاء الاصطناعي هي الألعاب التعاونية القابلة للملاحظة جزئياً. الألعاب دي بتطلب من اللاعبين إنهم يتعاونوا مع بعض، خصوصاً إن فيه جوانب في اللعبة قابلة للملاحظة بشكل جزئي بس، فهنا وكيل الذكاء الاصطناعي عنده كلمة مستهدفة، 'cold' (ساقع)، المستخدم ميعرفهاش. بس وكيل الذكاء الاصطناعي هيبعت إشارة، ولازم المستخدم يفسرها. ---
CHI-A48F7
CHI-A48F7-0110450.jpg
The user responds with a guess, and the AI agent sends another clue, and then at some point, the user might win the game if they guess correctly. We were really interested in investigating the direction of communication, because there are scenarios in which the human Must provide clues to the Ai agent, and those clues will differ than the clues that an Ai agent provides. In prior work, we were investigating interactions with just one model. We wanted to investigate whether social perceptions change and vary with different models. We had a total of six models, so for each of these models listed here, we had a guesser and a counter. We will refer to these models as model A, B, and C, because in our findings, we don't want to overgeneralize. If, for example, the users win more when they play with model B, that reinforcement learning is the best model for all human-AI
المستخدم بيرد بتخمين، ووكيل الذكاء الاصطناعي بيبعت إشارة تانية، وبعدين في مرحلة ما، المستخدم ممكن يكسب اللعبة لو خمن صح. كنا مهتمين جداً ببحث اتجاه التواصل، لأن فيه سيناريوهات بيبقى لازم فيها الإنسان هو اللي يقدم الإشارات لوكيل الذكاء الاصطناعي، والإشارات دي هتختلف عن الإشارات اللي بيقدمها وكيل الذكاء الاصطناعي. في شغلنا اللي فات، كنا بنبحث في التفاعلات مع نموذج واحد بس. كنا عايزين نبحث هل التصورات الاجتماعية بتتغير وتتنوع مع نماذج مختلفة. كان عندنا في المجمل ست نماذج، فلكل نموذج من النماذج اللي مكتوبة هنا، كان عندنا 'مخمن' و'معطي إشارات'. هنسمي النماذج دي نموذج 'أ' و'ب' و'ج'، لأننا في نتايجنا مش عايزين نعمم بزيادة. لو مثلاً المستخدمين كسبوا أكتر لما لعبوا مع نموذج 'ب'، ده مش معناه إن 'التعلم التعزيزي' (reinforcement learning) هو أحسن نموذج لكل تفاعلات ---
CHI-A48F7
CHI-A48F7-0122400.jpg
interactions. This is another example of gameplay. This is with model A. The target word is vanilla, and here you see the AI providing different clues, fruit,
البشر والذكاء الاصطناعي. ده مثال تاني لطريقة اللعب. ده مع نموذج 'أ'. الكلمة المستهدفة هي 'vanilla' (فانيليا)، وهنا بتشوف الذكاء الاصطناعي وهو بيقدم إشارات مختلفة، 'fruit' (فاكهة)، ---
CHI-A48F7
CHI-A48F7-0123450.jpg
chocolate, flavor,
'chocolate' (شوكولاتة)، 'flavor' (نكهة)، ---
CHI-A48F7
CHI-A48F7-0128450.jpg
bean. In another example, this is model B, and you
'bean' (حبة/بذرة). في مثال تاني، ده نموذج 'ب'، وهتلاحظ ---
CHI-A48F7
CHI-A48F7-0129500.jpg
see that the nature
إن طبيعة ---
CHI-A48F7
CHI-A48F7-0133450.jpg
of the, the nature of the clues are different, And this is model C, data
طبيعة الإشارات مختلفة، وده نموذج 'ج'، القايم على ---
CHI-A48F7
CHI-A48F7-0146450.jpg
driven. So we were really interested in how the direction of communication, various Ai agents and the perception of their partner's Ai identity affect social perceptions of their partner. and also impact gameplay outcome.
البيانات. فكنا مهتمين جداً بإزاي اتجاه التواصل، وتنوع وكلاء الذكاء الاصطناعي، وتصور هوية الشريك كذكاء اصطناعي بيأثروا على التصورات الاجتماعية للشريك، وكمان بيأثروا على نتيجة اللعب. ---
CHI-A48F7
CHI-A48F7-0167000.jpg
We ran an experiment in which participants were assigned to different conditions. They were told they were either playing against an Ai or a human. They played as a guesser or a giver, and they were assigned to any of the models I just described. They also played ten different games and then they took a survey on intelligence, rapport, likability, and then an Ai
عملنا تجربة اتوزع فيها المشاركين على حالات مختلفة. اتقال لهم إنهم بيلعبوا يا إما ضد ذكاء اصطناعي أو ضد إنسان. لعبوا كـ 'مخمن' أو 'معطي إشارات'، واتوزعوا على أي نموذج من اللي لسه وصفاهم. لعبوا كمان عشر ألعاب مختلفة وبعدين ملو استبيان عن الذكاء، والتآلف، والقبول، وبعدين ---
CHI-A48F7
CHI-A48F7-0184500.jpg
score. Our findings show that for one of the AI agents, model A, when participants believed to be playing against an AI, they found their partners to be more intelligent when they were playing the role of giver than when they were playing the role of guesser. Prior work has shown there to be bias against bots when they disclose their identity, but in our experiment,
نتيجة الذكاء الاصطناعي. نتايجنا بتبين إنه بالنسبة لواحد من وكلاء الذكاء الاصطناعي، وهو نموذج 'أ'، لما المشاركين اعتقدوا إنهم بيلعبوا ضد ذكاء اصطناعي، لقوا إن شركاءهم أذكى لما كانوا بيلعبوا دور 'معطي الإشارات' أكتر مما لما كانوا بيلعبوا دور 'المخمن'. الشغل اللي فات بين إن فيه انحياز ضد الـ 'بوتس' لما بيكشفوا عن هويتهم، بس في تجربتنا، ---
CHI-A48F7
CHI-A48F7-0218000.jpg
we observed the bias against the AI to be more present when users play the guesser versus the giver. One potential explanation for the differences in social perception could be explained around a user's feeling like they are in control when they play as the giver and not when they play as the guesser. When users played the role of guesser and believed their partners to be an AI, the model C agent is perceived more positively in likability and rapport than model A and model B. Of all three agents, model C has a higher average score and lower average number of
لاحظنا إن الانحياز ضد الذكاء الاصطناعي كان موجود أكتر لما المستخدمين بيلعبوا دور 'المخمن' مقابل 'معطي الإشارات'. تفسير محتمل للفروقات في التصور الاجتماعي ممكن يتشرح حولين إحساس المستخدم بإنه مسيطر لما بيلعب دور 'معطي الإشارات' ومسيطرش لما بيلعب دور 'المخمن'. لما المستخدمين لعبوا دور 'المخمن' واعتقدوا إن شركاءهم ذكاء اصطناعي، وكيل النموذج 'ج' اتصور بشكل إيجابي أكتر في القبول والتآلف من نموذج 'أ' ونموذج 'ب'. ومن بين التلات وكلاء، نموذج 'ج' عنده متوسط نتيجة أعلى ومتوسط عدد أدوار (turns) أقل من نموذج 'أ' ونموذج 'ب'. ---
CHI-A48F7
CHI-A48F7-0268450.jpg
turns than both model A and model B. One potential explanation for these results could be that it is a better model than the others, and it is better at giving clues that are interpretable by users. We also found that users took longer, more turns to win when interacting with a human and wanted to see why. We analyzed the clues that they gave for when users felt they were interacting with an AI or interacting with a human and found these themes. We found repetition when users thought they were interacting with a human. They would repeat words to kind of signal that it is an important word. We also found users were using relational words. So, for example, when the target word is cold, they would use hot and then use the word opposite to indicate that hot is the opposite of cold. And we also found suggestive words when users felt that they were interacting
تفسير محتمل للنتايج دي ممكن يكون إنه نموذج أحسن من الباقيين، وأحسن في تقديم إشارات يقدر المستخدمين يفسروها. لقينا كمان إن المستخدمين خدوا وقت أطول وأدوار أكتر عشان يكسبوا لما اتفاعلوا مع إنسان وكنا عايزين نعرف ليه. حللنا الإشارات اللي قدّموها لما المستخدمين حسوا إنهم بيتفاعلوا مع ذكاء اصطناعي أو بيتفاعلوا مع إنسان ولقينا الأنماط دي. لقينا تكرار لما المستخدمين افتكروا إنهم بيتفاعلوا مع إنسان؛ كانوا بيكرروا الكلمات كنوع من الإشارة لإنها كلمة مهمة. لقينا كمان إن المستخدمين كانوا بيستخدموا كلمات علائقية. فمثلاً، لما الكلمة المستهدفة تكون 'cold' (ساقع)، كانوا بيستخدموا 'hot' (سخن) وبعدين يستخدموا كلمة 'opposite' (عكس) عشان يشيروا لإن سخن هو عكس ساقع. ولقينا كمان كلمات إيحائية لما المستخدمين حسوا إنهم بيلعبوا ضد إنسان. ---
CHI-A48F7
CHI-A48F7-0314450.jpg
against a human. Our results highlight that it's not enough to evaluate social perceptions with one AI model, as the results across all of our AI agents varied. A lot of researchers in the CHI community and researchers who are investigating the relationships between humans and AI might think that it's enough to investigate with one AI agent, but our results seem to suggest that it's important to investigate with various AI agents. We also found that in human-AI interaction, the directionality of communication has an impact on the perception of an agent and is important to consider the implications of directionality when investigating human-AI interaction.
نتايجنا بتسلط الضوء على إنه مش كفاية نقيم التصورات الاجتماعية بنموذج ذكاء اصطناعي واحد، لأن النتايج عبر كل وكلاء الذكاء الاصطناعي بتوعنا كانت متفاوتة. باحثين كتير في مجتمع CHI والباحثين اللي بيبحثوا في العلاقات بين البشر والذكاء الاصطناعي ممكن يفتكروا إنه كفاية البحث مع وكيل ذكاء اصطناعي واحد، بس نتايجنا بتبين إن مهم البحث مع وكلاء ذكاء اصطناعي متنوعين. لقينا كمان إنه في التفاعل بين الإنسان والذكاء الاصطناعي، اتجاهية التواصل ليها تأثير على تصور الوكيل، ومهم وضع آثار الاتجاهية في الاعتبار عند بحث التفاعل بين الإنسان والذكاء الاصطناعي. ---
CHI-A48F7
CHI-A48F7-0323460.jpg
Thank you for listening and I'll be happy to take your questions.
شكراً لاستماعكم ويسعدني إني أجاوب على أسئلتكم.
CHI-AE83E
CHI-AE83E-0010000.jpg
Hello everyone, I am Dr. Sara Saveri and I would like to welcome you all to the paper presentation of UiSketch, a large-scale
مرحباً جميعاً، أنا دكتورة سارة سافيري وعاوزة أرحب بيكم كلكم في عرض ورقة البحث عن UiSketch، وهو مجموعة كبيرة من البيانات'
CHI-AE83E
CHI-AE83E-0033000.jpg
dataset of UI element sketches. UiSketch dataset is the first large-scale dataset of 17,979 hand-drawn sketches of 21 UI element categories collected from 967 participants, including UI UX designers, front-end developers, HCI and CS grad students from 10 different countries. Before delving into the details of this paper and what it contributes, I would like to establish some background
من رسومات عناصر واجهة المستخدم. UiSketch هو أول مجموعة بيانات كبيرة تشمل 17,979 رسم يدوي لـ 21 فئة من عناصر واجهات المستخدم تم جمعها من 967 مشارك، بما في ذلك مصممي واجهات المستخدم وتجربة المستخدم، مطوري الواجهة الأمامية، وطلاب الدراسات العليا في تخصصات HCI وعلوم الحاسوب من عشر دول مختلفة. قبل ما ندخل في تفاصيل الورقة دي وإسهاماتها، عاوزة أشرح الخلفية الخاصة بالموضوع'
CHI-AE83E
CHI-AE83E-0079000.jpg
on this area of research. Recent research, where HCI and AI interact, we require large-scale datasets to train AI models. In UI design research, currently, the two large datasets we have are Ricoh and Swire. Ricoh contains around 72,000 UI screenshots and Swire contains about 3,800 load-free screens. Swire was sketched by five designers and it is specific for the GUI similarity search system. Also, it does not have annotations for training any other supervised learning models. On the other hand, research in sketch recognition uses these other large datasets, so we wanted to create a large dataset for UI load-free prototypes and understand how humans and computers recognize UI element sketches. We approached the research following the footsteps of prior research in collecting drawings of
في المجال ده من البحث. في الأبحاث الحديثة، اللي بتتفاعل فيها الـHCI مع الذكاء الاصطناعي، لازم يكون عندنا مجموعات بيانات كبيرة لتدريب نماذج الذكاء الاصطناعي. في أبحاث تصميم واجهات المستخدم، حالياً، عندنا مجموعتين بيانات كبار هما Ricoh وSwire. Ricoh فيها حوالي 72,000 صورة لواجهات المستخدم وSwire فيها حوالي 3,800 واجهة خالية من التحميل. Swire كانت مرسومة من خمسة مصممين وهي مخصوصة لنظام البحث عن التشابه في واجهة المستخدم. كمان، ما فيهاش بيانات لتدريب نماذج التعلم بالإشراف. من ناحية تانية، الأبحاث في التعرف على الرسومات بتستخدم المجموعات الكبيرة الأخرى دي، وإحنا عاوزين نصمم مجموعة بيانات كبيرة للنماذج الأولية لواجهات المستخدم الخالية من التحميل ونفهم إزاي البشر والحواسيب بتتعرف على رسومات عناصر واجهة المستخدم. بدأنا البحث باتباع خطوات الأبحاث السابقة في جمع رسومات'
CHI-AE83E
CHI-AE83E-0099000.jpg
real-life objects. First, we identified a list of UI element categories that needed to be included in the dataset. Then, we created questionnaires and collected the respective sketches from participants. In both the paper and digital questionnaires, we show a sample UI element image on the left-hand side that we randomized to get varied sketches, even if the user traces it. On the right-hand side, we
أشياء الحياة الواقعية. أولاً، حددنا قائمة بفئات عناصر واجهة المستخدم اللي محتاج تضمينها في مجموعة البيانات. بعدين، صممنا استبيانات وجمعنا الرسومات المطلوبة من المشاركين. في الاستبيانات الورقية والرقمية، عرضنا صورة لعناصر واجهة المستخدم على الجانب الشمال وعشوائنا الصور للحصول على رسومات متنوعة، حتى لو المستخدم بيعمل تتبع للصورة. على الجانب اليمين،'
CHI-AE83E
CHI-AE83E-0124000.jpg
provide a sketching area. We used these questionnaires and over a span of three years, we collected UI element sketches from 967 participants from 10 different countries. Our participants included UI UX designers with 2 to 14 years of prior prototyping experience, front-end developers, HCI and CS grad students. These sketches were collected with pen and pencil using a paper questionnaire. However, in the digital questionnaires, we restricted them only to use
وفرنا مساحة للرسم. استخدمنا الاستبيانات دي وعلى مدار ثلاث سنين، جمعنا رسومات عناصر واجهة المستخدم من 967 مشارك من عشر دول مختلفة. المشاركين شملوا مصممي واجهات المستخدم وتجربة المستخدم اللي عندهم خبرة من سنتين لـ14 سنة في النمذجة السابقة، مطوري الواجهة الأمامية، وطلاب الدراسات العليا في تخصصات HCI وعلوم الحاسوب. الرسومات دي تم جمعها باستخدام القلم والمسطرة من خلال استبيانات ورقية. لكن في الاستبيانات الرقمية، قيدناهم باستخدام فقط'
CHI-AE83E
CHI-AE83E-0128000.jpg
styluses. Once we collected all these sketches, we started our data processing.
الأقلام الرقمية. وبعد ما جمعنا كل الرسومات دي، بدأنا معالجة البيانات.'
CHI-AE83E
CHI-AE83E-0144400.jpg
To this, we created a script that takes the paper questionnaire as input and uses contour detection to return labeled, close-cropped element sketches. We then manually verified these sketches. We removed badly drawn, contaminated, and irrelevant sketches to finalize this dataset.
للقيام بذلك، صممنا سكريبت بياخد الاستبيانات الورقية كمدخلات وبيستخدم الكشف عن الحواف لإرجاع رسومات لعناصر واجهة المستخدم مُصنفة ومقطوعة بدقة. بعد كده، راجعنا الرسومات دي يدوياً. شلنا الرسومات المشوهة أو الملوثة وغير المناسبة علشان ن finalized مجموعة البيانات دي.'
CHI-AE83E
CHI-AE83E-0156000.jpg
This dataset contains around 70 % of digital sketches and 30 % of paper sketches. We tried as much as possible to keep the classes balanced, so we have about 850 images for each
مجموعة البيانات دي تحتوي على حوالي 70% من الرسومات الرقمية و30% من الرسومات الورقية. حاولنا بقدر الإمكان نحافظ على التوازن بين الفئات، عشان عندنا حوالي 850 صورة لكل'
CHI-AE83E
CHI-AE83E-0176000.jpg
UI element category. With this dataset, we first wanted to analyze the drawing conventions and the differences between the sketches created by experienced designers and novices. The key difference was in the level of details added to the sketch. Experienced designers and developers tend to keep the sketches minimal, whereas novice designers such as HCI and CS grad students added more detail
فئة من فئات عناصر واجهة المستخدم. بمساعدة مجموعة البيانات دي، كنا عاوزين نحلل أولاً العادات المستخدمة في الرسم والاختلافات ما بين الرسومات المصممة من مصممين محترفين والرسومات اللي اتعملت من مبتدئين. الفرق الأساسي كان في مستوى التفاصيل اللي أُضيفت للرسم. المصممين والمطورين المحترفين بيميلوا يحتفظوا بالرسم مختصر، بينما المصممين المبتدئين زي طلاب الدراسات العليا في HCI وعلوم الحاسوب بيضيفوا تفاصيل أكتر'
CHI-AE83E
CHI-AE83E-0179000.jpg
to each sketch. To set a baseline, we conducted
لكل رسم. علشان نحدد نقطة البداية، عملنا'
CHI-AE83E
CHI-AE83E-0201400.jpg
another study with 76 UI UX designers where we showed them UI element sketches and asked them to identify their categories. Our results show that designers can identify Ui element sketches with 96 % accuracy. To understand how they identify the Ui elements. We looked at the Ui elements that they are mostly confused. We observed that human recognition errors are usually caused due to the confusion between semantically similar UI element categories.
دراسة تانية بمشاركة 76 مصمم واجهة مستخدم وتجربة مستخدم، عرضنا عليهم رسومات لعناصر واجهة المستخدم وطلبنا منهم تحديد الفئات الخاصة بيها. النتائج بتظهر إن المصممين قدروا يحددوا رسومات عناصر واجهة المستخدم بدقة 96%. علشان نفهم إزاي هم بيحددوا العناصر دي، نظرنا للعناصر اللي كانت بتسبب لهم ارتباك. لاحظنا إن الأخطاء اللي بيعملها البشر في التعرف عادة بتكون بسبب الخلط بين الفئات المشابهة من الناحية المعنوية لعناصر واجهة المستخدم.'
CHI-AE83E
CHI-AE83E-0216450.jpg
Moving on, we trained 26 different image and sketch classification neural networks and identified that current state-of-the-art Neural networks could identify Ui element sketches with around 80 to 90 % accuracy. The best performing model was Resnet 01:52 with
بعد كده، دربنا 26 شبكة عصبية مختلفة لتصنيف الصور والرسومات، ولقينا إن الشبكات العصبية الحالية المتقدمة قادرة على تحديد رسومات عناصر واجهة المستخدم بدقة حوالي 80% إلى 90%. أفضل نموذج كان ResNet-152 بدقة'
CHI-AE83E
CHI-AE83E-0221000.jpg
92 % accuracy. We further looked at how they identify these UI elements and
92%. بعدين، درسنا كيفية تحديدهم للعناصر دي و'
CHI-AE83E
CHI-AE83E-0248000.jpg
found that computational recognition errors are solely caused due to the confusion between the structural similarity of UI element categories. To further analyze the reasoning behind why the model confuses few UI element categories, we used its ResNet-152 model as a feature tracker along with TSNE to reduce the dimensionality of the test dataset images and plotted them as a 2D graph. This graph clearly shows that the computational classification model clusters the image based on the structural
اكتشفنا إن الأخطاء في التعرف الحسابي كانت كلها نتيجة الخلط بسبب التشابه البنيوي لفئات عناصر واجهة المستخدم. علشان ندرس أكتر الأسباب اللي بتخلي النموذج يخلط بين بعض الفئات، استخدمنا نموذج ResNet-152 كمتعقب للخصائص مع TSNE لتقليل الأبعاد للصور الموجودة في مجموعة الاختبار، ورسمناهم كرسم بياني ثنائي الأبعاد. الرسم ده بيوضح بوضوح إن النموذج الحسابي لتصنيف الصور بيجمع الصور بناءً على التشابه البنيوي'
CHI-AE83E
CHI-AE83E-0277000.jpg
similarities unlike humans. This dataset can be used for a lot of potential applications. Recently, a number of research projects attempted to automate transformation of LoFi sketches to code. As this dataset contains varied sketches, we can generate synthetic LoFi prototypes using the dataset and train such model. Another application would be to retrieve inspirational UI designs by sketching. This dataset could help train these AI models as well. Finally, we could also create sketch auto-completion to support designers while sketching LoFi
عكس البشر. مجموعة البيانات دي ممكن تُستخدم في تطبيقات كتير محتملة. مؤخراً، عدد من المشاريع البحثية حاولت تحول رسومات LoFi لكود تلقائيًا. بما إن مجموعة البيانات دي بتحتوي على رسومات متنوعة، ممكن نولد نماذج LoFi اصطناعية باستخدامها وندرب النموذج على كده. تطبيق تاني ممكن يكون استرجاع تصميمات واجهات المستخدم الإبداعية عن طريق الرسم. مجموعة البيانات دي ممكن تساعد في تدريب نماذج الذكاء الاصطناعي دي. وأخيراً، ممكن نصمم ميزة إكمال الرسومات تلقائيًا لدعم المصممين أثناء رسم تصاميم LoFi'
CHI-AE83E
CHI-AE83E-0291000.jpg
prototypes. We would also like to highlight a few limitations and how we want to address them in our future work. Currently, this dataset only covers smartphone platform. In the future, we would like to collect more sketches for other platforms as well. Similarly, we would like to collect sketches of LoFi
النماذج الأولية. عاوزين كمان نسلط الضوء على بعض القيود وإزاي بنخطط نعالجها في أبحاثنا المستقبلية. حالياً، مجموعة البيانات دي بتغطي بس منصة الهواتف الذكية. في المستقبل، عاوزين نجمع رسومات أكتر لمنصات تانية برضه. بنفس الطريقة، عاوزين نجمع رسومات لـLoFi'
CHI-ABB4F
CHI-ABB4F-0032470.jpg
Hi, my name is Marianela and I am going to present the paper Resisting the Medicalization of Menopause, Reclaiming the Body Through Design, on behalf of my co-authors Marie-Louise Jules Sondergaard and Madeleine Weiland from KTH Royal Institute of Technology in Sweden. In this paper, we address menopause as a social justice matter. We present a participatory design approach and we show how design can come close to the body to make space for experiences
مرحبًا، انا اسمي مارينيلا وانا هاقدم الورقة البحثية "مقاومة التوجه الطبي لانقطاع الطمث، استعادة الجسم عن طريق التصميم"، نيابة عن المؤلفين المشاركين ماريا-لويس جولز سونديرجارد ومادلين فايلاند من معهد KTH الملكي للتكنولوجيا في السويد. في الورقة دي، بنتناول انقطاع الطمث كموضوع متعلق بالعدالة الاجتماعية. بنقدم نهج التصميم التشاركي وبنوضح ازاي التصميم يقدر يقرب للجسم ويوفر مساحة للتجارب.'
CHI-ABB4F
CHI-ABB4F-0074450.jpg
of menopause. Menopause is a normal part of life, but a narrow view of health in the Western biomedical tradition contributed to frame this entire aspect of life as a problem, and in particular as a medical problem. Medicalizing menopause and tying it to gender has pathologized women's health and marginalized women and anyone who is not a cisgender man. For example, hormonal and menstrual variations in adolescents are considered normal, whereas in midlife they are framed as symptoms. Given the gender and age-based marginalization that people experience, we address menopause as a social justice
انقطاع الطمث. انقطاع الطمث جزء طبيعي من الحياة، لكن الرؤية الضيقة للصحة في التقاليد الطبية الغربية ساهمت في وضع هذا الجانب الكامل من الحياة كأنه مشكلة، وبالتحديد كأنه مشكلة طبية. اتجاه علاج انقطاع الطمث كمشكلة طبية وربطه بالنوع، قد أدى إلى تصنيف صحة النساء كمرض وتهميش النساء وأي شخص مش راجل متوافق مع النوع الاجتماعي. على سبيل المثال، يتم اعتبار التغيرات الهرمونية والدورة الشهرية في المراهقين شيء طبيعي، بينما في منتصف العمر يتم تصنيفها كأعراض. بالنظر للتهميش القائم على النوع والعمر اللي الناس بتعاني منه، بنتناول انقطاع الطمث كموضوع عدالة اجتماعية.'
CHI-ABB4F
CHI-ABB4F-0087455.jpg
matter. And as feminist design researchers, we want to foster societal change through design. So we first interviewed 12 women about their lived experiences of menopause, and our data analysis
اجتماعي. وبصفتنا باحثين في تصميم نسوي، بنهدف إلى تعزيز التغيير الاجتماعي من خلال التصميم. علشان كده، قمنا بإجراء مقابلات مع 12 امرأة عن تجاربهم في الحياة مع انقطاع الطمث، وتحليل البيانات اللي حصلنا عليها.'
CHI-ABB4F
CHI-ABB4F-0109450.jpg
highlights changes that are both bodily rooted and socially shaped. Participants have put effort, sometimes over years, into noticing and trying to make sense of these changes, and in developing strategies to live with and as a changing body, cultivating self-knowledge, acceptance, and trust in oneself.
بيسلط الضوء على التغييرات اللي جذورها جسدية والشكلها اجتماعي. المشاركات بذلوا مجهود، أحيانًا على مدار سنوات، علشان يلاحظوا ويحاولوا يفهموا التغييرات دي، ويطوروا استراتيجيات للتأقلم مع الجسم المتغير، وتنمية المعرفة الذاتية والقبول والثقة في النفس.'
CHI-ABB4F
CHI-ABB4F-0120455.jpg
These insights guided us in planning a participatory design workshop that we ran with three women and that invited them to slowly and gently engage with the
الرؤى دي ساعدتنا في تنظيم ورشة تصميم تشاركية عملناها مع ثلاث نساء ودعيناهم للمشاركة ببطء وبلطف مع'
CHI-ABB4F
CHI-ABB4F-0134450.jpg
body. Participants chose three experiences that they wanted to design for, bodily anxiety, introversion, and sleep is gold. They used a variety of materials to create two main design
الجسم. المشاركات اختاروا ثلاث تجارب كانوا عاوزين يصمموا لها، القلق الجسدي، الانطوائية، والنوم هو الذهب. استخدموا مجموعة متنوعة من المواد علشان يصمموا فكرتين رئيسيتين.'
CHI-ABB4F
CHI-ABB4F-0217455.jpg
concepts. The cocoon. This is a wearable that would be wrapped around the body, constraining it while allowing to push, and helping one breathe through an audio guidance and through mirroring movements. This addresses two opposite but coexisting intentions in participants' lives. When contracting inside the cocoon, they allow themselves to be introverts. When releasing and stretching out, they are making space. The second concept is the spike mat, a mat covered with sensors, actuators, that get perceived as spikes that massage the muscles. The mat would detect when the wearer is moving uncomfortably in bed, triggering the actuators with an intensity that avoids waking up the person. Here, participants knew that they were not going to solve their pain, and instead the mat would allow them to get pleasurable sensations on the skin and muscles. These concepts illustrate how design can engage with the body in a holistic way that does not isolate the body from the context in which it lives, while reclaiming its centrality. In contrast with the values proposed by a magicalized discourse, these concepts do not try to get rid of menopause experiences, nor to conceal, quantify, classify,
المفهوم الأول هو الشرنقة. دي قطعة قابلة للارتداء بتلف حوالين الجسم، بتضغط عليه وفي نفس الوقت بتساعده على التنفس من خلال إرشادات صوتية ومن خلال مرايا للحركة. المفهوم ده بيعالج نيتين متناقضتين لكن متعايشين في حياة المشاركات. لما بيكون الجسم منكمش جوا الشرنقة، بيتيح للمشاركات أن يكونوا انطوائيات. ولما بينطلقوا وبتمددوا، بيخصصوا مساحة. المفهوم التاني هو سجادة الشوك، سجادة مغطاة بمستشعرات ومحركات، بيتم إدراكها كأنها شوك بتدلك العضلات. السجادة بتكتشف لما المستخدم بيتحرك بطريقة غير مريحة في السرير، وبتشغل المحركات بدرجة مناسبة علشان تتجنب إيقاظ الشخص. هنا، المشاركات كانوا عارفين أنهم مش هيحلوا آلامهم، لكن بدل كده السجادة هتسمح لهم يحصلوا على إحساس ممتع على الجلد والعضلات. الأفكار دي بتوضح ازاي التصميم يقدر يتفاعل مع الجسم بطريقة شاملة من غير ما يعزل الجسم عن السياق اللي بيعيشه فيه، بينما بيستعيد مركزيته. على العكس من القيم اللي بيقترحها خطاب علاجي سحري، الأفكار دي مش بتحاول تتخلص من تجارب انقطاع الطمث، ولا تخفيها، ولا تصنفها، ولا تسيطر عليها.'
CHI-ABB4F
CHI-ABB4F-0265450.jpg
or control them. So based on our findings, we discuss in the paper three ways in which HCI can further address menopause as a social justice matter. First, by supporting resistance of a magicalized view. And here the key is that resisting magicalization is about dismantling altogether the role of people going through menopause as patients. Second, by being close to the ever-changing body. Here we emphasize that design should always question assumptions of an ideal, immutable body. And third, by including a plurality of bodies and experiences. And here in the paper we share our struggles to recruit a diverse set of people, and we recognize that more work is needed to really ensure
أو تتحكم فيها. بناءً على النتائج اللي وصلنا لها، ناقشنا في الورقة ثلاث طرق يقدر من خلالها مجال HCI يعالج انقطاع الطمث كقضية عدالة اجتماعية. أولًا، من خلال دعم مقاومة الخطاب العلاجي السحري. وهنا المفتاح هو إن مقاومة العلاج السحري بتتعلق بتفكيك دور الناس اللي بيمروا بانقطاع الطمث كمرضى. ثانيًا، من خلال القرب من الجسم المتغير باستمرار. هنا بنؤكد إن التصميم لازم دايمًا يسأل عن الافتراضات المتعلقة بالجسم المثالي والثابت. وثالثًا، من خلال تضمين تعددية الأجساد والتجارب. وهنا في الورقة شاركنا صعوباتنا في تجنيد مجموعة متنوعة من الناس، واعترفنا إن محتاجين شغل أكتر علشان نضمن بالفعل'
CHI-ABB4F
CHI-ABB4F-0275450.jpg
plurality. To conclude, we suggest designing with the premise that a willful acceptance of the body can be a motor for transforming menopause
التعددية. علشان نختم، بنقترح التصميم بناءً على فرضية إن القبول الإرادي للجسم ممكن يكون دافع لتحويل تجارب انقطاع الطمث.'
CHI-ABB4F
CHI-ABB4F-0299470.jpg
experiences. And we see many opportunities for HCI to shift from technology that avoids the body by quantifying it, controlling it, or denying its experiences, towards technology that elevates the ever-changing body as a step towards societal change. We thank all the people that made this research possible, and thank you for watching.
التجارب. وبنرى فرص كتيرة لمجال HCI إنه يتحول من التكنولوجيا اللي بتتجنب الجسم عن طريق تصنيفه أو التحكم فيه أو إنكار تجاربه، إلى تكنولوجيا بترفع من قيمة الجسم المتغير باستمرار كخطوة نحو التغيير المجتمعي. بنشكر كل الناس اللي ساهموا في جعل البحث ده ممكن، وشكرًا لكم على المشاهدة.'
CHI-26CA3
CHI-26CA3-0014450.jpg
Hello, I'm Kai Lukoff, a Ph.D. candidate at the University of Washington. Today I'm presenting our research on how the design of YouTube influences user sense of agency.
أهلاً، أنا كاي لوكوف، طالب دكتوراه في جامعة واشنطن. النهاردة أنا بقدم البحث بتاعنا عن إزاي تصميم اليوتيوب بيأثر على إحساس المستخدمين بالتحكم.
CHI-26CA3
CHI-26CA3-0024000.jpg
For smartphone apps, a common business model is to capture and resell attention. This leads designers to devise features that capture as much of that
بالنسبة لتطبيقات الموبايل، نموذج العمل المنتشر هو جذب الانتباه وإعادة بيعه، وده بيخلي المصممين يبتكروا مميزات تجذب أكبر قدر من الانتباه ده.
CHI-26CA3
CHI-26CA3-0047000.jpg
attention as possible. In everyday life, people attempt to resist all kinds of temptations, such as smoking, drinking, and spending money. However, the temptation with the highest rate of failure of all is media use, which today increasingly happens in mobile apps. And of course, this high failure rate is not an accident when those mobile apps are designed to capture as much
في حياتنا اليومية، الناس بتحاول تقاوم كل أنواع الإغراءات، زي التدخين والشرب وصرف الفلوس. لكن، الإغراء اللي نسبة الفشل فيه هي الأعلى على الإطلاق هو استخدام الميديا، واللي النهاردة بيحصل بشكل متزايد في تطبيقات الموبايل. وطبعاً، نسبة الفشل العالية دي مش صدفة لما تطبيقات الموبايل دي بتكون مصممة عشان تستحوذ على أكبر قدر ممكن من الانتباه.
CHI-26CA3
CHI-26CA3-0090000.jpg
attention as possible. Concerned design practitioners and researchers have responded from within what we call the screen time paradigm. In this paradigm, the objective is to limit the time that users spend on their phone. Yet users report that these tools often indiscriminately block access to needed features, and that reducing screen time is a poor proxy for their actual goals. In this research, we instead explore how to support the user's sense of agency, that is, an individual's experience of being the initiator of their actions in the world. A lack of agency over technology use is linked to negative life impacts, such as a loss of productivity, sleep, and social
الانتباه قدر الإمكان. ممارسي التصميم والباحثين المهتمين استجابوا من داخل اللي بنسميه "نموذج وقت الشاشة". في النموذج ده، الهدف هو تحديد الوقت اللي المستخدمين بيقضوه على تليفوناتهم. لكن المستخدمين بيبلغوا إن الأدوات دي غالبًا بتحجب الوصول بشكل عشوائي للميزات الضرورية، وإن تقليل وقت الشاشة ده بديل ضعيف لأهدافهم الحقيقية. في البحث ده، إحنا بدلًا من كده بنستكشف إزاي ندعم إحساس المستخدم بالقدرة على التصرف، يعني، تجربة الفرد في إنه يكون هو اللي بيبدأ أفعاله في العالم. نقص الإحساس بالقدرة على التصرف في استخدام التكنولوجيا مرتبط بتأثيرات سلبية على الحياة، زي فقدان الإنتاجية والنوم والعلاقات الاجتماعية.
CHI-26CA3
CHI-26CA3-0113450.jpg
relationships. A second characteristic of the screen time paradigm is that it relies upon external features that apply universally to all apps. In this research, we instead shift the focus to redesigning the internal features that might cause an app to be problematic in the first place. As a test case for this approach, we focus on YouTube, one of the most popular social media apps in the
العلاقات. خاصية تانية لنموذج وقت الشاشة هي إنه بيعتمد على خصائص خارجية بتنطبق بشكل عام على كل التطبيقات. في البحث ده، إحنا بدل كده بنحوّل التركيز لإعادة تصميم الخصائص الداخلية اللي ممكن تخلي التطبيق يسبب مشكلة من الأساس. وكمثال عملي للنهج ده، إحنا بنركز على يوتيوب، واحد من أشهر تطبيقات السوشيال ميديا في الـ
CHI-26CA3
CHI-26CA3-0125000.jpg
world. Our first research question asks, which existing features of the YouTube mobile app influence sense of agency? To investigate this question, we conducted a survey with 120 YouTube
. سؤال البحث الأول بتاعنا بيسأل: إيه الميزات الموجودة حالياً في تطبيق اليوتيوب للموبايل اللي بتأثر على الإحساس بالفاعلية؟ عشان نتحرى السؤال ده، عملنا استبيان مع 120 يوتيوب
CHI-26CA3
CHI-26CA3-0135000.jpg
users from the U.S. Our survey asked, what are three things about the mobile app that lead you to feel least and most in control over how you spend your time on YouTube?
استبياننا سأل مستخدمين من أمريكا: إيه هما التلات حاجات في تطبيق الموبايل اللي بتخليك تحس إن عندك أقل وأكتر سيطرة على طريقة قضائك لوقتك على يوتيوب؟
CHI-26CA3
CHI-26CA3-0160450.jpg
This diverging bar chart shows the results. They are ordered according to which design feature was mentioned the most frequently by respondents. The features that primarily led participants to feel least in control were recommendations, ads, and autoplay. The features that primarily led participants to feel most in control were playlists, search, subscriptions, watch history, and play
الرسم البياني الشريطي المتفرع ده بيورِّي النتايج. مترتبين على حسب خاصية التصميم اللي اتذكرت أكتر حاجة من المشاركين. الخصائص اللي خلت المشاركين بشكل أساسي يحسوا إن تحكمهم كان أقل حاجة هي التوصيات والإعلانات والتشغيل التلقائي. والخصائص اللي خلت المشاركين بشكل أساسي يحسوا إن تحكمهم كان أكتر حاجة هي قوايم التشغيل والبحث والاشتراكات وتاريخ المشاهدة والتشغيل.
CHI-26CA3
CHI-26CA3-0186450.jpg
controls. Let's look at one example. Survey participants reported that playlists led them to feel more control. One participant wrote, I can create playlists or queue videos in advance to limit what I watch to a specific list instead of endlessly searching around for what I want. So as compared to searching around or browsing recommendations, playlists had a clear end that served as a signal for them to finish their session on YouTube.
تحكم. خلينا نشوف مثال واحد. المشاركين في الاستبيان قالوا إن قوائم التشغيل خلتهم يحسوا بتحكم أكتر. واحد من المشاركين كتب: "أقدر أعمل قوائم تشغيل أو أرتب الفيديوهات مقدماً عشان أحدد اللي بتفرج عليه لقائمة معينة بدل ما أدور بشكل لا نهائي على اللي أنا عايزه." عشان كده مقارنةً بالتدوير أو تصفح الاقتراحات، قوائم التشغيل كان ليها نهاية واضحة ودي كانت إشارة ليهم عشان ينهوا جلستهم على يوتيوب.
CHI-26CA3
CHI-26CA3-0216450.jpg
Our second research question asks, what changes to these internal features might increase sense of agency? We were particularly interested in understanding how to change features that the survey participants said sometimes made them feel more in control and sometimes made them feel less in control. To investigate this question, we conducted co-design sessions with 13 general users of the YouTube mobile app. This included an exercise in which participants sketched and explained what changes would give them greater control.
سؤالنا البحثي التاني بيسأل، إيه التغييرات في الميزات الداخلية دي اللي ممكن تزود إحساسهم بالتحكم؟ كنا مهتمين بشكل خاص إننا نفهم إزاي نغير الميزات اللي المشاركين في الاستطلاع قالوا إنها ساعات بتخليهم يحسوا إنهم مسيطرين أكتر وساعات بتخليهم يحسوا إنهم مسيطرين أقل. عشان نبحث السؤال ده، عملنا جلسات تصميم مشترك مع 13 مستخدم عادي لتطبيق يوتيوب على الموبايل. ده شمل تمرين المشاركين فيه رسموا وشرحوا إيه التغييرات اللي ممكن تديهم تحكم أكبر.
CHI-26CA3
CHI-26CA3-0242000.jpg
Participants explained that they used Youtube in two different use cases and that they preferred different levels of control for each. When they had a specific intention in mind, such as watching a tutorial on how to solve a Rubik's cube, they wanted more control. When they did not have a specific intention in mind and they just wanted to relax or felt bored, they wanted to let YouTube take control. This leads us to our design
المشاركون وضحوا إنهم كانوا بيستخدموا يوتيوب في حالتين استخدام مختلفتين، وإنهم كانوا بيفضلوا مستويات مختلفة من التحكم لكل حالة. لما كان بيكون عندهم نية معينة في دماغهم، زي مشاهدة فيديو تعليمي عن كيفية حل مكعب روبيك، كانوا عايزين تحكم أكبر. ولما مكنش عندهم نية معينة في دماغهم، وكانوا بس عايزين يسترخوا أو حسوا بالملل، كانوا بيفضلوا يسيبوا يوتيوب هو اللي يتحكم. وده بيوصلنا لتصميمنا.
CHI-26CA3
CHI-26CA3-0267000.jpg
implications. Our first design implication is to adapt the interface to the specificity of the user's intention. For example, when the user has a specific intention, the designer can present a search-first interface. On the other hand, when the user has no specific intention, the designer can present a recommendations-first interface. Switching between these two interfaces might be done either manually by the user or
تداعيات. أول تداعيات تصميمية لينا هي إننا نكيف الواجهة عشان تتناسب مع طبيعة نية المستخدم. على سبيل المثال، لما يكون عند المستخدم نية محددة، المصمم ممكن يقدم واجهة البحث أولاً. من ناحية تانية، لما المستخدم ميكونش عنده نية محددة، المصمم ممكن يقدم واجهة التوصيات أولاً. التبديل بين الواجهتين دول ممكن يتم بشكل يدوي عن طريق المستخدم أو
CHI-26CA3
CHI-26CA3-0288450.jpg
automatically by the system. Our second design implication is to help the user make a micro plan, which we define as a lightweight plan that guides behavior for a short time, usually just a single session of use. So, for example, the playlists feature in YouTube helped with this, but we argue that there's more designers can do here by drawing upon knowledge from the field of behavior planning.
تلقائيًا عن طريق النظام. تاني تأثير تصميمي لينا هو مساعدة المستخدم إنه يعمل خطة مصغرة، واللي بنعرفها بإنها خطة خفيفة بتوجه السلوك لمدة قصيرة، عادةً بتكون مجرد جلسة استخدام واحدة. فمثلًا، خاصية قوايم التشغيل في يوتيوب ساعدت في ده، لكننا بنرى إن المصممين ممكن يعملوا أكتر هنا عن طريق الاعتماد على المعرفة من مجال تخطيط السلوك.
CHI-26CA3
CHI-26CA3-0303000.jpg
In conclusion, we encourage designers to think beyond the screen time paradigm, support greater control when users have a specific intention in mind, and help users micro plan. Thank you.
في الختام، بنشجع المصممين إنهم يفكروا أبعد من مفهوم وقت الشاشة، ويدعموا تحكم أكبر لما يكون المستخدمين عندهم نية محددة في بالهم، ويساعدوا المستخدمين على التخطيط المصغر. شكراً.
CHI-903ED
CHI-903ED-0008450.jpg
Hello everyone, I'm Sarah and today I present our concept drive
أهلاً بيكم كلكم، أنا سارة والنهاردة بقدم الفكرة التصورية بتاعتنا.
CHI-903ED
CHI-903ED-0059450.jpg
you on behalf of my co-authors Ahmed Chams, Robin Piening, Jomna Abdelrahman and Florian Alt. So today users are surrounded by ubiquitous devices that collect data about them. To give you a few examples, imagine some daily life scenarios like a rental apartment, a museum or a train station, which are places that might be unfamiliar to you. At the same time, data collection might happen in places that are familiar to you, such as a friend's place, your office shared kitchen or your daily way to work. Note that we also picked these examples to cover private, semi-public and public spaces. However, users are oftentimes unaware of this data collection. To address this in such and also other scenarios, we build
أنا بكلمكم نيابة عن زملائي المؤلفين المشاركين أحمد شمس، روبن بينينج، جومانا عبد الرحمن وفلوريان ألت. فالنهاردة، المستخدمين محاطين بأجهزة منتشرة في كل مكان، بتجمع عنهم بيانات. عشان أديكوا كام مثال، تخيلوا كام سيناريو من حياتنا اليومية زي شقة إيجار، متحف، أو محطة قطر، ودي أماكن ممكن ما تكونوش مألوفة ليكم. وفي نفس الوقت، جمع البيانات ده ممكن يحصل في أماكن مألوفة ليكم، زي بيت صاحبك، المطبخ المشترك في مكتبك، أو طريقك اليومي للشغل. خدوا بالكم إننا كمان اخترنا الأمثلة دي عشان نغطي الأماكن الخاصة، وشبه العامة، والعامة. لكن، المستخدمين في أغلب الأوقات مبيبقوش واخدين بالهم من جمع البيانات ده. عشان نتعامل مع ده في السيناريوهات دي، وفي سيناريوهات تانية كمان، إحنا بنينا
CHI-903ED
CHI-903ED-0066450.jpg
PriView. And with PriView, we can visualize potential privacy intrusion by ubiquitous devices.
براي فيو. ومع براي فيو، نقدر نتصور انتهاك محتمل للخصوصية بواسطة الأجهزة المنتشرة في كل مكان.
CHI-903ED
CHI-903ED-0071450.jpg
In this paper, we present two prototypes on two output
في الورقة البحثية دي، بنقدم نموذجين أوليين على مخرجين.
CHI-903ED
CHI-903ED-0118450.jpg
devices, namely a handheld mobile application and a hands-free head-mounted display. We investigated both in a user study. For the mobile application, we implemented five different visualizations and participants conducted search tasks in our lab using these visualizations. The visualizations were covering various information about devices, for example, bounding boxes just framing the device, text labels with additional information, 3D shapes indicating the device's tracking space, sensor icons, and then also highlighting thermal hotspots on the device using the thermal camera.
أجهزة، وبالتحديد تطبيق موبايل يدوي وشاشة عرض بتتلِبس على الراس بدون استخدام الأيدي. درسنا الاتنين دول في دراسة للمستخدمين. بالنسبة لتطبيق الموبايل، نفذنا خمس تصورات مختلفة والمشاركين عملوا مهام بحث في معملنا باستخدام التصورات دي. التصورات دي كانت بتغطي معلومات مختلفة عن الأجهزة، زي مثلاً مربعات حدودية بتحدد الجهاز بس، وتسميات نصية بمعلومات إضافية، وأشكال ثلاثية الأبعاد بتوضح مساحة تتبع الجهاز، وأيقونات سنسور، وكمان بتبرز النقط الساخنة حرارياً على الجهاز باستخدام الكاميرا الحرارية.
CHI-903ED
CHI-903ED-0182450.jpg
And then we built a second prototype in a head-mounted display, which served two purposes. On one hand, it is another modality that is hands-free, but it also allowed us to implement the sample scenarios that I showed earlier using virtual reality. For the head-mounted display, we again implemented five different visualizations and showed them to participants in all six sample scenarios. These captures here are taken from the rental apartment scene. Most of the visualizations were similar to the mobile application, so we again had these bounding boxes. text labels, and 3D shapes indicating the device's tracking space. But we also had two that were specific to the HMD, namely this permanent warning indicator here, and then also we had some floor markers, again indicating the area where users would be tracked by the device. In these scenarios, we did not have a concrete task for participants, but they could explore the environment
وبعد كده بنينا بروتوتايب تاني في نضارة رأسية، واللي كان ليها غرضين. من ناحية، هي طريقة عرض تانية بتشتغل من غير إيدين، لكن كمان سمحت لنا ننفذ السيناريوهات التجريبية اللي وريتها قبل كده باستخدام الواقع الافتراضي. بالنسبة للنضارة الرأسية، رجعنا تاني طبقنا خمس أشكال عرض مختلفة ووريناهم للمشاركين في الست سيناريوهات التجريبية كلها. اللقطات دي هنا متاخدة من مشهد شقة الإيجار. أغلب أشكال العرض كانت شبه تطبيق الموبايل، عشان كده رجعنا استخدمنا صناديق تحديد، وملصقات نصية، وأشكال ثلاثية الأبعاد بتوضح مساحة تتبع الجهاز. لكن كان عندنا اتنين كمان كانوا خاصين بالنضارة الرأسية، وتحديدا مؤشر التحذير الدايم ده هنا، وبعد كده كمان كان عندنا علامات أرضية، تاني بتوضح المنطقة اللي الجهاز هيتبع فيها المستخدمين. في السيناريوهات دي، مكنش عندنا مهمة محددة للمشاركين، لكن كان ممكن يستكشفوا البيئة.
CHI-903ED
CHI-903ED-0189450.jpg
freely. So that was both of our applications. And I'd now like to give you a quick overview of our results.
بحرية. فكده دول كانوا التطبيقين بتوعنا. ودلوقتي أحب أديكوا نبذة سريعة عن نتايجنا.
CHI-903ED
CHI-903ED-0202450.jpg
We found that participants overall liked our concept. They found it interesting, easy to use, and also fun. And also they stated that it could actually help to increase their privacy awareness.
لقينا إن المشتركين بشكل عام عجبهم المفهوم بتاعنا. لاقوه شيق، وسهل في الاستخدام، وممتع كمان. وكمان قالوا إنه ممكن فعلاً يساعد في زيادة وعيهم بالخصوصية.
CHI-903ED
CHI-903ED-0251450.jpg
We also found that the The amount of information of course needs to be balanced with visual overload, so for instance, in places where participants would already expect data collection. they would expect the Pryview to provide less details other than in places where they would not expect data collection. Also they wished for it to be interactive. So a common suggestion was to have this kind of permanent indicator, but then provide details on demand. Lastly, Pryview should not only protect the privacy of the that person using the application, but also others' privacy. So for instance, participants were concerned that their personal information could be revealed to others using Pryview.
ولقينا كمان إن كمية المعلومات طبعاً لازم تكون متوازنة مع العبء البصري الزايد، فمثلاً، في الأماكن اللي المشاركين بيكونوا متوقعين فيها جمع بيانات، كانوا بيتوقعوا إن Pryview يقدم تفاصيل أقل من الأماكن اللي ميكونوش متوقعين فيها جمع بيانات. وكمان كانوا بيتمنوا إنه يكون تفاعلي. فكان اقتراح شائع هو إنه يكون فيه مؤشر دائم بالشكل ده، لكن يوفر التفاصيل عند الطلب. أخيراً، Pryview مش بس المفروض يحمي خصوصية الشخص اللي بيستخدم التطبيق، لكن كمان خصوصية الآخرين. فمثلاً، المشاركين كانوا قلقانين إن معلوماتهم الشخصية ممكن تتكشف للآخرين اللي بيستخدموا Pryview.
CHI-903ED
CHI-903ED-0260450.jpg
To summarize, we built two sample implementations for our concept Pryview, namely one handheld mobile application, and one
باختصار، بنينا تطبيقين تجريبيين لمفهومنا Pryview، وهما تطبيق موبايل محمول، وواحد
CHI-903ED
CHI-903ED-0298450.jpg
mobile application. And one hands-free application in a head-mounted display. We implemented various visualizations for each of these prototypes. We found that Pryview can increase privacy awareness in various use cases, that balancing the right amount of information and visual overload is highly context-dependent, that Pryview should provide mechanisms to interact, and lastly, that Pryview needs to consider individual privacy needs and potentially provide mechanisms to opt out being shown to others. For more details, I'd like to refer you to our paper.
تطبيق موبايل. وتطبيق تاني بيشتغل بدون استخدام الأيدي في شاشة رأسية. نفذنا طرق عرض مختلفة لكل نموذج أولي من دول. لقينا إن برايفيو ممكن يزود الوعي بالخصوصية في حالات استخدام مختلفة، وإن الموازنة بين الكمية المناسبة من المعلومات والحمل البصري الزيادة بيعتمد بشكل كبير على السياق، وإن برايفيو المفروض يوفر آليات للتفاعل، وأخيراً، إن برايفيو محتاج ياخد في الاعتبار احتياجات الخصوصية الفردية وممكن يوفر آليات لعدم الظهور للآخرين. لمزيد من التفاصيل، ممكن ترجعوا للبحث بتاعنا.
CHI-903ED
CHI-903ED-0303450.jpg
With that, I'd like to thank you for your attention and I'm looking forward to the Q&A session.
وبكده، أحب أشكركم على اهتمامكم، وأتطلع لفقرة الأسئلة والأجوبة.
CHI-11811
CHI-11811-0014450.jpg
Hi everyone, my name is Karthik and I'm a PhD student in the DGP lab at the University of Toronto. Today I'm presenting our work on task allocation for human-robot collaboration.
أهلاً بيكم يا جماعة، إسمي كارثيك وأنا طالب دكتوراه في معمل DGP في جامعة تورنتو. النهاردة أنا بقدم شغلنا عن توزيع المهام للتعاون بين الإنسان والروبوت.
CHI-11811
CHI-11811-0028450.jpg
In human-robot collaboration, humans and robots work together to achieve a common goal. One of the many challenges in HRC is determining what actions each member of the team should take, a problem known as task allocation.
في التعاون بين الإنسان والروبوت، بيشتغل الإنسان والروبوت سوا عشان يحققوا هدف مشترك. وواحد من التحديات الكتير في التعاون ده هو تحديد الإجراءات اللي كل فرد في الفريق المفروض ياخدها، وهي مشكلة معروفة باسم "توزيع المهام".
CHI-11811
CHI-11811-0050450.jpg
Prior work has proposed many solutions to task allocation. For instance, using offline or online planning methods which prescribe what the human and the robot should work on during the task. However, sometimes it is hard to pre-plan all aspects of an HRC before the task commences, and humans may want to make changes to the task allocations in an ad hoc fashion.
الشغل اللي فات قدم حلول كتير لتوزيع المهام. مثلاً، باستخدام طرق تخطيط أوفلاين أو أونلاين اللي بتحدد إيه اللي المفروض الإنسان والروبوت يشتغلوا عليه خلال المهمة. لكن، ساعات بيبقى صعب نخطط مسبقاً لكل جوانب التعاون بين الإنسان والروبوت قبل ما المهمة تبدأ، وممكن البشر يحبوا يعملوا تغييرات في توزيع المهام بشكل عفوي.
CHI-11811
CHI-11811-0067450.jpg
Hence, prior work has proposed behaviors that the robot can use to facilitate ad hoc collaboration, e.g. the robot can respond to user instructions in a reactive manner, or the robot can infer allocations such as by picking up objects near its gripper in a proactive manner.
ولذلك، الأبحاث السابقة اقترحت سلوكيات ممكن الروبوت يستخدمها عشان يسهل التعاون الفوري. على سبيل المثال، الروبوت ممكن يستجيب لتعليمات المستخدم بشكل تفاعلي، أو الروبوت ممكن يستنتج المطلوب منه، زي إنه يلقط حاجات قريبة من المشبك بتاعه بشكل استباقي.
CHI-11811
CHI-11811-0098450.jpg
We take inspiration from ad hoc Hrc and coordination behaviours exhibited by human teams, such as the notion of territoriality. and proxemics, and HCI work on designing interactions with proxemics in mind. Tabletop territoriality describes how collaborators distribute resources in a shared workspace while working on a task. Proxemic interaction states that we can exploit physical and spatial relationships with objects to enable fluent and natural interactions with
إحنا بنستلهم من الـ Hrc الـ ad hoc وسلوكيات التنسيق اللي بتظهرها الفرق البشرية، زي مفهوم الإقليمية والبروكسيمكس، وشغل الـ HCI في تصميم التفاعلات مع أخذ البروكسيمكس في الاعتبار. إقليمية أسطح العمل بتوصف إزاي المتعاونين بيوزعوا الموارد في مساحة عمل مشتركة وهما شغالين على مهمة. التفاعل البروكسيمكسي بيقول إننا ممكن نستغل العلاقات الفيزيائية والمكانية مع الأشياء عشان نتيح تفاعلات سلسة وطبيعية مع
CHI-11811
CHI-11811-0108450.jpg
them. In our work we make two contributions, a design space of novel task allocation techniques for ad hoc HRC, and a comprehensive evaluation of the design space.
هما. في شغلنا بنقدم إسهامين، مساحة تصميم لتقنيات مبتكرة لتوزيع المهام لـ HRC المخصص، وتقييم شامل لمساحة التصميم.
CHI-11811
CHI-11811-0120450.jpg
We propose three new explicit techniques where allocation occurs through deliberate user action, and three new implicit techniques where the robot allocates objects to itself based on simple heuristics.
بنقترح تلات تقنيات جديدة صريحة، التخصيص فيها بيتم عن طريق فعل متعمد من المستخدم، وتلات تقنيات جديدة ضمنية، الروبوت بيخصص فيها الحاجات لنفسه بناءً على قواعد استدلال بسيطة.
CHI-11811
CHI-11811-0142450.jpg
Our first technique is called menu, where the user can touch or point at an object to bring up an allocation menu. The second technique is called subtle relocation, through which the user can push an object towards the robot to assign it.
أول تقنية عندنا اسمها "menu"، واللي فيها المستخدم يقدر يلمس أو يشاور على كائن عشان تطلعله قائمة تخصيص. التقنية التانية اسمها "إزاحة خفية"، اللي من خلالها المستخدم يقدر يزق كائن ناحية الروبوت عشان يخصصه.
CHI-11811
CHI-11811-0156450.jpg
The last technique is called fixed territories, where the tabletop is divided into territories belonging to the user, the robot, and the group. Objects change allocation based on where they're located on the table.
آخر تكنيك بيتسمى "المناطق الثابتة"، هنا الترابيزة بتكون متقسمة لمناطق ملك للمستخدم والروبوت والمجموعة. والأشياء بيتغير تخصيصها حسب مكانها على الترابيزة.
CHI-11811
CHI-11811-0169450.jpg
Our first implicit technique is called distance, where the robot allocates objects based on how far away they are from its base. The.
أول تقنية ضمنية لينا اسمها المسافة، الروبوت بيخصص فيها الأجسام على أساس مدى بُعدها عن القاعدة بتاعته.
CHI-11811
CHI-11811-0186450.jpg
The second technique is called gaze, where user gaze triggers territories, which the robot avoids picking up objects from, And robot territories which it actively picks up objects from, and?
التقنية التانية اسمها النظر، حيث نظرة المستخدم بتحفز مناطق، الروبوت بيتجنب إنه يشيل منها حاجات، ومناطق خاصة بالروبوت هو بيشيل منها حاجات بنشاط، و؟
CHI-11811
CHI-11811-0202450.jpg
And the last implicit technique, which is called proximity, where territories are formed when the user or the robot picks up objects from a region. Subsequently, the robot avoids picking up objects from the user's territory and only picks up objects from its own territory.
وآخر أسلوب ضمني، اللي بيُطلق عليه التقارب، حيث بتتكون مناطق لما المستخدم أو الروبوت بيلتقطوا أجسام من منطقة. وبعد كده، الروبوت بيتجنب يلتقط أجسام من منطقة المستخدم، وبيلتقط أجسام من منطقته هو بس.
CHI-11811
CHI-11811-0211450.jpg
We evaluated the techniques on a set of block stacking tasks, where the user manipulated yellow blocks and the robot manipulated black blocks.
قيّمنا التقنيات على مجموعة من مهام رص المكعبات، حيث المستخدم حرّك المكعبات الصفرا والروبوت حرّك المكعبات السودة.
CHI-11811
CHI-11811-0231450.jpg
We studied coupled tasks, where the user and the robot had to place blocks alternatively to complete a structure, And decoupled tasks where the user and the robot could independently complete their own structures. We also studied object placement. Objects for the user and the robot could be scattered around the workspace, or they could be pre-sorted.
إحنا درسنا مهام مقترنة، حيث كان على المستخدم والروبوت إنهم يحطوا مكعبات بالتناوب عشان يكملوا هيكل. ومهام غير مقترنة، حيث كان ممكن للمستخدم والروبوت إنهم يكملوا الهياكل بتاعتهم بشكل مستقل. إحنا كمان درسنا وضع الأشياء. الأشياء الخاصة بالمستخدم والروبوت ممكن تكون متناثرة في مساحة العمل، أو ممكن تكون مترتبة مسبقًا.
CHI-11811
CHI-11811-0256450.jpg
Here are some clips from the user study. We found that implicit techniques led to faster task completion times,
دي شوية مقاطع من دراسة المستخدمين. لقينا إن التقنيات الضمنية أدت لوقت أقل في إنجاز المهام،
CHI-11811
CHI-11811-0263450.jpg
led to increased concurrent activity or simultaneous time spent by the user and the robot on the task,
أدى إلى زيادة النشاط المتزامن أو قضاء وقت متزامن من المستخدم والروبوت على المهمة.
CHI-11811
CHI-11811-0267450.jpg
led to increased robot errors or incorrect
أدت لزيادة أخطاء الروبوت أو غلط
CHI-11811
CHI-11811-0280450.jpg
allocations, led to an increased perception of being able to parallelize or work on the task and allocate objects simultaneously, and finally led to a reduced perception of being in control over task allocation.
التخصيصات، أدت إلى إدراك متزايد للقدرة على التوازي أو الشغل على المهمة وتخصيص العناصر في نفس الوقت، وأدت في النهاية إلى إدراك أقل بالتحكم في تخصيص المهام.
CHI-11811
CHI-11811-0300450.jpg
Based on these findings, we think that future task allocation could be improved by building implicit allocation techniques with explicit control mechanisms. For example, in the proximity technique, the user's territory is implicitly formed when they pick up objects in the region, but the user could set the territory size explicitly.
بناءً على النتائج دي، احنا شايفين إن ممكن تحسين توزيع المهام في المستقبل عن طريق بناء أساليب تخصيص ضمنية مع آليات تحكم صريحة. على سبيل المثال، في تقنية التقارب، منطقة المستخدم بتتشكل ضمنياً لما ياخد حاجات في المنطقة، لكن المستخدم ممكن يحدد حجم المنطقة دي بشكل صريح.
CHI-11811
CHI-11811-0302450.jpg
Thank you very much for your attention.
شكرًا جدًا على اهتمامك.
CHI-5B402
CHI-5B402-0014450.jpg
Hello, I'm delighted to share with you our most recent research, which engages people with the sounds of the wild and features this guy, an eastern bristlebird.
أهلاً، أنا مبسوط إني أشارككم أحدث بحث لينا، اللي بيربط الناس بأصوات الحياة البرية وبيبرز الكائن ده، طائر أبو شوكة الشرقي.
CHI-5B402
CHI-5B402-0054450.jpg
Eastern bristlebirds are rare and they're hard to see, and the population where I live in Queensland, Australia, lives among these tall tussocks grasses, which makes them really hard to find. We can find them, however, if we know how to listen for them. I spent two years learning calls using recordings of captive birds first, so you can imagine how excited I was when I first heard their calls in the wild. That's the picture you see here. And I'd like to introduce you to one of those calls. Isn't it beautiful?
طيور الشوكة الشرقية نادرة وصعب رؤيتها، والتجمعات اللي عايشة في المكان اللي أنا فيه في كوينزلاند بأستراليا، بتعيش وسط حشائش التيسوك العالية دي، وده بيخلّي العثور عليهم صعب جداً. لكن ممكن نلاقيهم لو عرفنا إزاي نسمع صوتهم. أنا قضيت سنتين بتعلّم نداءاتهم باستخدام تسجيلات لطيور محبوسة الأول، فممكن تتخيلوا كنت متحمس قد إيه لما سمعت نداءاتهم لأول مرة في البرية. دي الصورة اللي انتوا شايفينها هنا. وأحب أسمّعكوا واحدة من النداءات دي. مش جميلة؟
CHI-5B402
CHI-5B402-0072450.jpg
Music to my ears. Well, audio recordings are pretty easy to gather, as long as you can get to that rugged habitat. But making sense of the audio after you have it is the real challenge. We don't have a lot of knowledge about the calls of bristlebirds, which makes it extra
موسيقى في وداني. طب، التسجيلات الصوتية سهلة أوي إنك تجمعها، طول ما تقدر توصل للبيئة الوعرة دي. بس إنك تفهم الصوت ده بعد ما تحصل عليه، هو ده التحدي الحقيقي. معرفتنا قليلة عن نداءات طيور البريستل بيرد، وده بيخليه زيادة
CHI-5B402
CHI-5B402-0131450.jpg
difficult. I'd like to introduce you to another sound. Now when we get audio recordings, the fantastic benefit is that we have visualizations of sound as well. Have a look at the shape in that what we call a spectrogram and see if you can recognize when I play it again what that call sounds like and looks like. So we informed the design of the Bristle Whistle Challenge by our prior work in the lab. We designed and play tested a prototype as part of this paper. It includes mixed activities, including gameful, meaning structured and goal oriented tasks, and playful, meaning free form and exploratory
صعب. أنا حابب أعرّفكوا على صوت تاني. دلوقتي لما بنجيب تسجيلات صوتية، الميزة الرهيبة هي إن بيكون عندنا تصوّرات مرئية للصوت كمان. بصّوا على الشكل في اللي بنسمّيه spectrogram وشوفوا لو هتقدروا تميّزوا لما أشغّلها تاني النداء ده بيكون عامل إزاي وشكله إزاي. عشان كده، إحنا استندنا في تصميم تحدي صفارة الشعيرات على شغلنا اللي فات في المعمل. إحنا صمّمنا نموذج أولي وعملنا له اختبارات لعب كجزء من الورقة البحثية دي. بيتضمّن أنشطة متنوّعة، منها أنشطة ألعابية (gameful) يعني مهام منظّمة وموجّهة نحو هدف، وأنشطة لعب حر (playful) يعني شكلها حر واستكشافية.
CHI-5B402
CHI-5B402-0141450.jpg
activities. This was intended to inform design of nature engagement and conservation focused citizen science.
أنشطة. وده كان مقصود بيه توجيه تصميم علوم المواطنين اللي بتركز على مشاركة الناس في الطبيعة والحفاظ عليها.
CHI-5B402
CHI-5B402-0164450.jpg
One of the features of the Bristle Whistle Challenge was that participants were receiving guidance by a conversational agent. As you see here highlighted, that was Bob the Bristle Bird. He gave various comments throughout the entire challenge to assist people, and this provided sense making and emotional support.
من مميزات تحدي صفارة بريستل إن المشاركين كانوا بيتلقوا إرشادات عن طريق مساعد محادثة. زي ما أنتم شايفين هنا متظلل، ده كان بوب طائر بريستل. هو كان بيقدم تعليقات متنوعة على مدار التحدي كله عشان يساعد الناس، وده كان بيوفر مساعدة على فهم الأمور ودعم عاطفي.
CHI-5B402
CHI-5B402-0174450.jpg
The gamefully matching of media was actually quite engaging for people when they were looking at spectrograms and audio
مطابقة الوسائط بشكل مسلي كانت في الحقيقة شيقة جداً للناس لما كانوا بيشوفوا الأطياف الصوتية والصوت.
CHI-5B402
CHI-5B402-0185450.jpg
or matching the different photographs to specific calls of particular species. It helped to hone in attention on the calls of the
أو مطابقة الصور المختلفة بأصوات محددة لأنواع معينة. ده ساعد على تركيز الانتباه على أصوات الـ
CHI-5B402
CHI-5B402-0198450.jpg
birds. The playfully exploring media included looking at photographs and spectrograms and listening to the calls as well, but without any particular tasks.
الطيور. الوسائل اللي استخدموها في الاستكشاف بطريقة مرحة كانت بتشمل الفرجة على صور ومخططات طيفية وسماع الأصوات كمان، بس من غير أي مهام معينة.
CHI-5B402
CHI-5B402-0217450.jpg
It also included imagining what people might see as creative shapes in the spectrograms of calls. Do you see that flower shape in there like I do? Well, these types of activities piqued interest and elicited creativity in our
وكمان كانت بتشمل تخيل الأشكال الإبداعية اللي الناس ممكن تشوفها في الرسوم الطيفية (السبكتروجرامات) بتاعة الأصوات. شايف شكل الوردة ده زي ما أنا شايف؟ أهو، الأنواع دي من الأنشطة أثارت الاهتمام وطلعت الإبداع في الـ
CHI-5B402
CHI-5B402-0229450.jpg
participants. Then we explored how they might apply what they learned to quest for calls that they were meant to identify in unfamiliar spectrograms.
وبعدين، شفنا إزاي يقدروا يطبقوا اللي اتعلموه عشان يدوروا على نداءات كان مطلوب منهم يتعرفوا عليها في رسوم طيفية مش مألوفة.
CHI-5B402
CHI-5B402-0242450.jpg
proved to be very useful, providing them all that variety of context ahead of time for identifying particular birds. And they enjoyed the experiences.
طلع مفيد أوي، ووفر لهم كل التنوع ده من السياق مقدماً عشان يحددوا طيور معينة. واستمتعوا بالتجارب دي.
CHI-5B402
CHI-5B402-0247450.jpg
It also revealed sensory and animal preference differences,
وكشفت كمان عن فروقات في التفضيلات الحسية والحيوانية،
End of preview. Expand in Data Studio
README.md exists but content is empty.
Downloads last month
36