F ACM

Having trouble with OKX Wallet - 'gas fee error' error while trying to transfer funds. Any solution? Wallet phrase: notice leisure umbrella over east select damage pink devote drip federal august

this guy's the GOAT

❤❤

AAPKA VIDEO BAHUT ACHA LAGA HAMA HAM BHI EAK YOUTUB CHAINAL CREAT KIYA HAI HAMARA BHI HELP KARNA BHI PLESE 🙏🙏🙏🙏

We know how papers are published: it is all about collegiality, being nice, and if necessary, neglecting standards and scientific/academic integrity. I have seen poor (impossible may be a better word) configurations touted as successes.

Why mono sound?(

Really nice article highlighting some persistent issues with modern AI/ML research like cherrypicking, leakage and overfitting. This is a problem I see both in published work and internal projects at companies. It requires a lot of careful testing across a broad variety of datasets to ensure that models with a large number of parameters work well. In this particular instance, it is ironic that good old SA can outperform RL and generalizes better.

🐻‍❄️✨

thank Igor for trying to return back the dignity of computer architecture research. It is been a long time that non-sense papers are published with poor data and implementation.

if the seller never intended to sell the legitimate file, he can corrupt the file then hash it and send it... it doesn't solve the problem that because you cannot preview the content of the file... but that's another question for a different day

Bro, I've just started a rust course today (nov 2024) and already found this treasure talk hidden in KZbin... great talk, we need this dude back for an updated version of this talk

Next is the Venus score!

God_Bliss; in an instant I like it_love it

Spent a while looking into the paper/article, and frankly this is the most ridiculous pile of crap I have ever seen. I feel bad for the Google research team responsible for this paper, speculation is fine but unfounded speculation that basically sums up to "it doesn't look like it should work to me specifically, so it's fraud and Google is corrupt" is seriously dumb.

Contact with A.I. 🛸

Niklaus REALLY didn't like C/C++ 🤣 Now, let's rebuild a Lilith, just for kicks !

Fantastic presentation! Your approach in ‘Contrast then Memorize: Semantic Neighbor Retrieval-Enhanced Inductive Multimodal Knowledge Graph Completion’ is incredibly innovative and well-executed. I’m currently working on a similar topic, and your insights on combining contrastive learning with semantic neighbor retrieval for unseen entities are invaluable. Would you be open to sharing some advice on how you optimized the contrastive loss in your framework? I'm particularly interested in how you handled retrieval efficiency and the role of pre-trained vision-language models. Looking forward to applying some of these concepts to my own research. Amazing work once again, and I’d love to hear more about your future projects!

Where is the song?!

1:12:43

It was the fantastic session 😊👍

Wonderful conversation.

Thanks.

Kavon Cape

This is a positive step toward inclusive research. However, the slides are unfortunately difficult to read.

11.53

Nice

Good talk!

so fantastic !!!

C´mon. Let´s get real. Don Knuth is the man. Period.

Koch Ramp

Cardelli, know him of seminal Object Orientation formal semantics papers, how versatile of him to also move in this domain.

Darlene Villages

The great man, Lastei

🎯 Key points for quick navigation: 00:00:04 *🎓 ACM's tech talk is part of a commitment to lifelong learning, supporting a large global membership of computing professionals and students.* 00:01:00 *📚 ACM offers resources to enhance skills and career opportunities for its members, helping them stay competitive in the field of computing.* 00:01:28 *🌐 ACM provides members with timely computing information, digital library access, international conferences, and supports education and research.* 00:02:12 *🔄 ACM members can use technology to solve critical problems and improve society, as part of the ACM's mission.* 00:03:06 *🔔 This webcast will be recorded; attendees will be notified via email when it's available in the ACM archives.* 00:03:49 *📢 Viewers can interact via social media widgets and take part in community discussions post-webcast.* 00:05:41 *🧩 Jeannette Wing introduces data science and emphasizes the importance of responsible data use for societal benefit.* 00:06:06 *📊 Data is generated by people, instruments, and sensors, collected, processed, stored, and analyzed for valuable insights.* 00:07:13 *🔍 Data analysis and visualization are crucial in making sense of data and conveying findings effectively.* 00:08:07 *⚖️ Ethical and privacy considerations are integral throughout the data lifecycle.* 00:08:33 *💡 Data science is defined as extracting value from data, where "value" is interpreted differently by end-users like scientists, policymakers, or business owners.* 00:11:19 *🌍 Data should be used to tackle grand societal challenges, while ensuring the responsible use of data addresses privacy and ethics.* 00:12:07 *🎯 Columbia University has a highly interdisciplinary data science institute aiming to advance data science and apply it broadly across fields.* 00:14:21 *🤝 The Data Science Institute collaborates extensively with industry and academia to foster data science research and innovation.* 00:19:17 *🎥 An example of applying causal reasoning in data science: analyzing the impact of movie casting choices on revenue.* 00:27:31 *🚀 Neural networks are being used in astrophysics to improve the analysis of cosmic imagery data.* 00:29:05 *📊 Data analysis of labor markets like Amazon Mechanical Turk shows they function more as monopsonies rather than traditional markets.* 00:30:08 *🤖 Automated financial advisors can learn investor preferences using reinforcement learning.* 00:31:28 *📜 Applying topic modeling to declassified diplomatic documents helps identify historical events, like the Saigon evacuation.* 00:33:17 *🔍 Responsible data use focuses on fairness, accountability, transparency, and ethics in AI, including robustness and explainability.* 00:34:40 *📊 Traditional and emerging trustworthiness properties are essential, such as safety, security, privacy, and usability.* 00:36:23 *🔧 The Deep Explorer system finds vulnerabilities in AI by using neuron coverage analogous to code coverage.* 00:37:52 *🛡️ Differential privacy techniques enhance DNNs' robustness by adding noise layers.* 00:39:07 *🌍 Columbia's Lamont center uses Pangeo, an open-source climate science analysis platform.* 00:41:01 *🌾 Climate predictions urge India to diversify grains due to changing monsoon patterns impacting rice reliance.* 00:42:06 *🏥 Columbia's Odyssey project aggregates over 500 million patient records for healthcare insights.* 00:44:08 *🔄 Treatment uniqueness is notable, with a quarter of hypertension treatments globally being unique.* 00:45:11 *💊 Observational data reveals genetic disposition differences, such as in Japanese diabetes treatments.* 00:47:43 *🧪 New data science methods in causal inference enhance healthcare outcomes, showing the value of observational data.* 00:49:26 *📈 Post-analysis verification enhances model trustworthiness, starting from model building stages.* 00:54:37 *🔍 Business commitment to responsible data use is crucial for maintaining customer trust and satisfaction.* 00:58:29 *⚖️ Causality modeling, particularly robust synthetic control, offers alternatives to randomized control trials.* 01:01:48 *⚖️ Legal and ethical considerations arise in data scraping practices for journalistic purposes.* 01:01:39 *🤔 Tension exists between collecting data for good and adhering to legal or policy restrictions.* 01:02:32 *🔍 The term "responsible data scraping" could address the balance between ethical data use and legal compliance.* 01:03:00 *👏 The session concluded, thanking Jeannette Wing for her informative presentation and attendees for their participation.* 01:03:29 *📅 The recorded talk will be made available online soon, with notifications for those registered.* Made with HARPA AI

Hello professor, your lectures is excellent, can you please share this book with me

VR headsets needs this with eye tracking,

Phenomenal❤

Fantastic explanation of real-world cryptojacking! What’s the best way to detect cryptojacking early?

Great insights! How does MinerRay stay ahead of evolving cryptojacking techniques?

You look 🔥 talking tech stuff!

Thank you for your humble honestly.

More people should make the effort to understand what you do/have done.

Wonderful Interview

Is there a recording for 2024 event?

"fire in their belly and want to change the world"

This video is highly underrated

The stars I wish to meet and study my PHD under plus Andrew Ng and Raffaello D'Andrea

error : Author are not tagged/Authors and Year must tagged for Author Year References references; Firstname Surename and have the style Authgroup applied to the entire name. --> what is that for? meanwhile i choose code 1 (number format instead of author year)