Sabitlenmiş Tweet
Are you wondering how you can evaluate some of the risks of a foundation model before you deploy it? Read on .... ibm.com/new/announceme…
English
Michael Hind
831 posts
Michael Hind
@michael_hind
IBM Distinguished RSM, working AI transparency, governance, explainability, and fairness. Proud husband & dad, Soccer lover. Tweets are my own.
Cortlandt Manor, NY Katılım Ağustos 2012
366 Takip Edilen440 Takipçiler
Michael Hind retweetledi
🚀 Just launched: BenchmarkCards – a toolkit for safer, more transparent #AI!
Built by @Lucy_Institute & @IBMResearch, supported by @NotreDameEthics, it’s now part of IBM Research’s open-source Risk Atlas Nexus for AI governance.
Learn more: tinyurl.com/2mecwezs
#AIDay
English
@peacock Can you please tell your Premiere league directors that we watch live soccer to ... watch live soccer!!! We should never miss *any* live action because of a replay.
When the ball is in play, we don't need crowd shots, coach shots, or replays.
*All* action is important!
English
Michael Hind retweetledi
Our team open-sourced two new models: Granite Guardian 3.0 2B and Granite Guardian 3.0 8B.
Read more: linkedin.com/pulse/ibm-open…
Hugging Face: huggingface.co/collections/ib…
Documentation: ibm.com/granite/docs/m…
Try them out!
English
Michael Hind retweetledi
We are thrilled to have co-authored this report offering ✨guidance on #AI transparency✨, with an superb set of coauthors @michael_hind @mmitchell_ai @stoyanoj @jennwvaughan @EmilyNexus @mcmillan_majora @kaschm @SarahWNewman + more
Shorenstein Center@ShorensteinCtr
We have just published the *CLeAR Documentation Framework for #AI Transparency and governance,* co-authored by experts across academia, industry, and civil society. The report shares guiding principles for immediate use in AI documentation: shorensteincenter.org/clear-document…
English
It's great to work with such amazing and impactful colleagues!
Kush Varshney कुश वार्ष्णेय@krvarshney
I wanted to share a bunch of ideas the human-centered and trustworthy AI teams at IBM Research labs worldwide have been simmering and are now externalizing. I encourage topics that might not be hyped, but are nevertheless important and ones that researchers believe in. /1
English
I enjoyed my recent interview on the AI Risk Reward podcast with host Alec Crawford (@alec06830).
You can hear it here: podcasts.apple.com/us/podcast/tru…
English
Michael Hind retweetledi
Can #AI be trusted?
Watch Dr Kush R. Varshney(@@krvarshney) explaining what is trustworthy AI and how it can be ensured!
Institute for Governance, Policies & Politics@IGPPVMF
#IGPPExpertTalks AI Special Series- Trusting the #Machines: Exploring the Reliability of #ArtificialIntelligence Ms. Heena Goswami (@HeenaAsks), editorial consultant, @IGPPVMF in conversation with Dr. Kush R. Varshney (@krvarshney ), Research Scientist, Senior Manager at IBM Thomas J. Watson Research Center(@IBMResearch) and author of '#Trustworthy Machine Learning.' 🎙️ Join Dr. Varshney as he unravels the world of Trustworthy #MachineLearning! Get ready for #insightful discussions on: 1️⃣ Can #AI be trusted? 2️⃣ How can #techexperts contribute to building reliable #AIsystems? 3️⃣ Exploring the #fairness and #explainability of #AImodels. 4️⃣ Journey towards #unbiasedAI - Is it complex or achievable? 5️⃣ What approach should #governments adopt in #regulatingAI? Don't miss this exclusive interview series, diving deep into the core of #AIreliability and #responsibility. #TrustworthyAI #AIInnovation #TechResponsibility #AITrust #AIFairness #ExplainableAI #TechInnovation #AIEthics #BiasInAI #TechChallenges #TechControl #AIInnovation #AIPolicy #Regulation #AIEthics #ResponsibleAI
English
@KateKayeReports @krvarshney @sharongoldman @thepamdixon @privacyforum @scheidegger @geomblog (That should read 1000s have used it.)
English
@KateKayeReports @krvarshney @sharongoldman @thepamdixon @privacyforum @scheidegger @geomblog AIF360 provides "speedometers" and "brakes" for AI bias. Based on the work of the scientific community, it contains tools to measure bias and techniques like Disparate Impact Remover to reduce/mitigate it. These tools are open source & free to anyone to use. 100s have used it.
English
I spoke to @thepamdixon & @KateKayeReports at the @privacyforum about their new report that found many AI governance tools meant to tackle bias and explainability turn out to be problematic or ineffective.
Read here: ⬇️
venturebeat.com/ai/many-ai-gov…
English
@KateKayeReports @krvarshney @sharongoldman @thepamdixon @privacyforum @scheidegger @geomblog The AIF360 guidance page (#guidance" target="_blank" rel="nofollow noopener">aif360.res.ibm.com/resources#guid…) provides guidance on its use and the complexity of this area. The tool does not advocate any bias threshold policy as seems to be implied by the article.
English
@KateKayeReports @krvarshney @sharongoldman @thepamdixon @privacyforum @scheidegger @geomblog Thus, I think it is misguided to disparage AIF360 based on a paper it cites that took a reasonable approach to an open problem: what is an acceptable threshold for bias.
English
Michael Hind retweetledi
“Modern advertising techniques use all available user information — including their privacy decisions — to construct more accurate user profiles. Users should be aware that their cookie decisions can have counterintuitive effects.” sciencefocus.com/news/declining… openreview.net/pdf?id=dFtpRph…
English
Michael Hind retweetledi
Interesting day learning about new #ArtificalIntelligence developments with @michael_hind of @IBM & @derektranchia at the @nassautract workshop held at @HofstraU- always something new to learn!
East Garden City, NY 🇺🇸 English
Michael Hind retweetledi
At the Understanding AI conference with @amurray71 . Great learning from presenters @michael_hind and @DerekTranchina at the @nassautract @HofstraU #ArtificialIntelligence #tlchat
English