Remote

Upol Ehsan

@upol@hci.social

mastodon 4.7.3

🎯 Making AI systems explainable & responsible so that people not at the table do not end up on the menu.

🛠️ HCI + AI + Philosophy to craft Human-centered Explainable AI (HCXAI).

💼 Georgia Tech, Data & Society | ex- {Google, IBM Research, Microsoft Research}

☕ Bubble Tea Connoisseur, Manchester United fan

#AI #ML #eXplainableAI #XAI #ResponsibleAI #AIEthics #HCI #HumanCenteredDesign #HumanCenteredExplainableAI #HCXAI #UX

0 Followers

0 Following

31 Posts

Joined November 02, 2022

Twitter:

https://twitter.com/UpolEhsan

LinkedIn:

https://www.linkedin.com/in/uehsan/

Google Scholar:

https://scholar.google.com/citations?hl=en&user=yKRlxFcAAAAJ

Upol Ehsan @upol@hci.social

· 22mo ago

Grant applications:
Convince us your idea is so groundbreaking it borders on sorcery—in 500 words. Then explain exactly how you'll pull off this miracle with no money, no staff, and a timeline shorter than a fruit fly's lifespan—in 300 words. Bonus points if you don’t cry.

#academia #academicmastodon

hci.social

🌱 hci.social

Upol Ehsan @upol@hci.social

· 25mo ago

Profs: How did you come up with your lab's name? Tell us the origin story.

What are some best practices? What are things to avoid?

Got a killer acronym? How did you come up with it?

Pls repost/share to get this post get out of the awful filter bubble

#academia #academicchatter #stem #science

hci.social

🌱 hci.social

Upol Ehsan @upol@hci.social

· 23mo ago

Me sh*tposting = many views and likes
Me carefully curating a near perfect paper explainer = nothing
Lesson = next paper explainer will be in the form of sh*tposting.

Upol Ehsan @upol@hci.social

· 22mo ago

I asked a senior colleague about grants.
Them: grant writing is basically a polite way of saying 'academic creative science fiction.' 😅

Upol Ehsan @upol@hci.social

· 22mo ago

Optimist: Glass half full.
Pessimist: Glass half empty.
AI researchers: A transformer is all you need.

Upol Ehsan @upol@hci.social

· 23mo ago

PSA for anyone trying to travel to Costa Rica for #CSCW2024 with a green card that has a notice of extension (I-797C): the embassy's website has contradictory information.

Here's the truth: You need a visa b/c Costa Rica does NOT recognize the extension letter.

Do NOT risk deportation.

hci.social

Upol Ehsan @upol@hci.social

· 22mo ago

Academic email threads: 'Please remove me from this chain.'
Next email: 'Please stop replying all.'
Next email: 'Replying all to agree—please stop replying all.'

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [13/n] Practical tips 💡 Think sociotechnically—seams often appear where social and technical factors intersect. 💡 Common seam hotspots: Data context shifts, model drifts, policy changes, factors beyond AI's training data or algorithmic scope.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [14/n] 💌 On a personal note, my biggest joy was working with @hal @qveraliao Samir Passi @Riedl. It felt like playing in a band where everyone is on song ⚒️ Years in the making, this project has been tough. Translating the abstract concept of seams into applied AI was anything but trivial.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [15/15] 🤗 Immensely grateful to the participants, Microsoft Research's FATE group, the loan officers for crafting scenarios, & Matthew Chalmers for discussing seamful design with me (fun fact: his work was my intro to seamful design). Paper link: https://arxiv.org/pdf/2211.06753 #academia #AI #HCI

arxiv.org

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [6/n] 🧩 Seams in AI= mismatches between dev assumptions & real-world use. Think of them as cracks where AI stumbles in the real world-- AI's blind spots. 📚 Examples of seams: data context shift (trained on US data, deployed in Bangladesh), policy shift, regulatory changes, etc.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [10/n] Step 2️⃣: Anticipating, locating, and crafting seams • Find the seam = what could cause the breakdowns? • Tether it = where are they located in the AI's lifecycle? • Craft it = what's the gap between expectation and reality? Step 3️⃣: Filtering relevant seams to enhance explainability and user agency • Filter relevant seams = which seams do we show, which do we hide? = • Justify why seams are shown = how does each improve actionability, contestability, appropriation?

Upol Ehsan @upol@hci.social

· 25mo ago

UPDATE: anon form -- https://docs.google.com/forms/d/e/1FAIpQLSeFzB14GLcLndgyuyjFco_oggU88yG5fI6nMulgxSLRPRCWPg/viewform?usp=sf_link

CS Profs: help out new faculty members. What are the mistakes you made when it came to PhD student recruitment when you started as a prof? (Note, the question is asking about recruitment tactics that you wish you had used when you started, not certain specific hiring decisions about people. Focus is on principles, not persons)

Please repost and help this get out of my filter bubble.

#AcademicChatter #academia #computerscience

docs.google.com

Upol Ehsan @upol@hci.social

· 22mo ago

Replying to

@mpt this is a really actionable input. Thanks for sharing this. My teammate who is in love with radar charts will appreciate it.

Upol Ehsan @upol@hci.social

· 22mo ago

Quiet quitting in academia: the fine art of swapping weekends for 'just a few hours' of work that last until Sunday night.

Upol Ehsan @upol@hci.social

· 23mo ago

Pls RT and help me get the word out:

Okay I bit the bullet and signed up for BlueSky-- if you're connected with me here, it will be great to connect there as well.

I pray this migration is better than the last one and we don't go back after 6 months.

https://bsky.app/profile/upolehsan.bsky.social

bsky.app

Upol Ehsan @upol@hci.social

· 23mo ago

Dear Viz folks, can you help with this radar chart issue where not all dimensions have the same scale? #HCI #viz #visualization

hci.social

🌱 hci.social

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

@luis_in_brief ah how interesting!

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [2/n] Explainable AI (XAI) Implications: ⚡️ Seams offer 'peripheral vision' of factors beyond the algorithm & reveal AI’s blind spots. 💡 This helps users unpack the 'why-not'—vital for handling AI failures. 🎯 Seamful XAI extends traditional XAI from the 'why' to the 'why-not'

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [3/n] Responsible AI (RAI) implications: ✅ Tackles gaps in current RAI methods—good at spotting risks, but weak on action. ✨ Our process empowers users to foresee AI harms, uncover respective seams, locate them in the AI lifecycle, & leverage them to meet user goals.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [4/n] Impact: 🚀 Incorporated into NIST's AI Risk Management Framework, a globally adopted RAI framework 🏢 Adopted by major companies for GenAI/LLM red-teaming. 🥇 First work to operationalize seamful design in (X)AI Here's a real case that illustrates why this matters ⤵️

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [5/n] 🏦 Real case: AI denies Ahmed's loan despite his great record. Why? 🕵️ A hidden seam: AI uses old 3-loan limit, unaware of new 5-loan policy. Ahmed has 4. 💡 If Nadia knew this mismatch, she could contest the AI's decision. ❓These mismatches are seams, let's define them.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [7/n] 🎯 The essence of seamful design (roots in Ubicomp): don’t just reveal seams, leverage them. But why? 💪 To augment user agency. ⚡️ For XAI, agency is operationalized as actionability, contestability, appropriation. 🤔 Seamful design is great, but there’s a challenge…

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [8/n] 🚧 Here's the challenge: Seams aren't explicit, not easy to find, and using them is hard. 🧭 People need a process. 🛠️ That's where our methodological contribution comes in – a design process to help people find + leverage seams.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [9/n] 📐 The Seamful XAI design process has 3 steps 1️⃣ Envisioning AI oopsies (breakdowns) 2️⃣ Anticipating and locating seams 3️⃣ Filtering seams to enhance explainability and user agency Let's break these down ⤵️ Step 1️⃣: Envisioning harms– here, think like a supervillain! (Yes, really!) • Ask: what could go wrong? • Play a supervillain and make the breakdowns happen! This adversarial thinking helps anticipate potential issues.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [11/n] But is the process any good? 🔬 We evaluated this process through a scenario-based design interview study with 43 participants across 9 countries and 6 domains. Here’re the key findings ⤵️ 🎖️ The design process was teachable and transferrable—everyone could craft seams and apply them to their fields. 🥳 The surprise hit? Roleplaying as a 'supervillain' made it fun and engaging! 🌟 But the biggest takeaway was learning how to leverage seams.

Upol Ehsan @upol@hci.social

· 23mo ago

Replying to

🧵 [12/n] Findings: 💪 Seamful XAI empowers users by expanding their options—giving them a voice, unlike seamless AI’s ‘take it or leave it’ model. 🚀 The process shifts AI design from reactive patching to a proactive one, empowering users to handle fallouts from AI failures better

Upol Ehsan @upol@hci.social

· 47mo ago

Is there any critical scholarship on #federated information architecture?

I am looking for pointers on the intersection of techno utopianism and notions of decentralized tech and how that intersects with privilege and power.

Here, privilege + power could also be geographical/political/infrastructural.

e.g., having the money to pay for servers, your country allowing payments in foreign currencies to be able to afford said servers.

#academicChatter #academicMastodon