Ali Tajer
@satajer
Followers
305
Following
2K
Media
7
Statuses
274
Professor @RPI (past Columbia & Princeton). Information Theory, Machine Learning, Statistical Inference
Troy, NY
Joined July 2015
RT @AlexGDimakis: Many people think high school level math problems means easy problems. Here is one from the recent IMO that current front….
0
1
0
Indeed. The more relevant question is who can monetize these models. IBM has significant legacy enterprise products ….
Actually, don't underestimate IBM! . The race is not just about training the best models. It's also about capturing the value of AI throughout the industry.
0
0
7
RT @nalidoust: I am a son of a former political prisoner in Iran and the nephew of an executed dissident and another political refugee in….
0
17
0
Except that the main role of an advisor is helping with posing the right questions, for which these tools are not too useful!.
At this point, you can use any AIs (perplexity, chatgpt, etc) as a PhD advisor on whatever topic you're getting deep into. It's pretty good. I have my preference. The core point: Research advice is no longer an elite academic university thing.
0
0
2
than to challenge them. So, while science has a significant role in providing ideas and evidence, its most effective role’d be discussing the tradeoffs in policies without prescribing one.
0
0
0
the tendency to process info in ways that support one’s pre-existing beliefs/identities. Unlike other reasoning errors, myside bias is independent of analytical/cognitives skills, and in fact those more seasoned with it are more likely to use it to defend their views rather ⬇️.
1
0
0
This is a great post. Complementing @boazbaraktcs’s rationale, I’m also reminded of the “myside bias” coined by Stanovich in. The Bias That Divides Us: The Science and Politics of Myside Thinking. arguing that social polarization is driven less by irrationality and more by ⬇️.
1
0
1
RT @boazbaraktcs: @littmath It’s a little known fact that many of the advances in AI started off when Arnold Schwarzenegger was spotted wal….
0
13
0
Paying back to the society that’s been my intellectual home. Will be happy to visit and give a talk on causality, representation learning, bandits, or their interactions.
0
0
8
AAAI 2025 Tutorial on Causal Representation Learning (February 26). If interested, please join us for an overview of Causal Representation Learning (CRL). More info re: content:. For more info. please contact one of us (Burak, Emre, Karthik, or me).
0
1
8
First time we all met in person after 4+ years of weekly online meetings to discuss causality and bandits! #NeurIPS2024!
1
1
44
And finally this super cool work — intervention-based learning of mixtures of DAGs, which are more powerful in abstracting the complexities of causal systems. We the usual suspects in the results: necessary & sufficient conditions for mixture recovery and an algorithm.
0
0
0
In addition to Burak’s work, Zirui will also present how to drop knowing the graph topology altogether — a major challenge in causal bandits, which has been understood for only do interventions. Zirui’s work addresses the general soft interventions setting #causality #bandits.
1
0
2
Burak & Emre will be presenting two papers on causal representation learning: (1) finite-sample guarantees (identifiability & algorithmic) and (2) unknown, multi-target intervention. Further evidence that our score-based method is highly versatile apt for diverse settings.
I'll be at #NeurIPS2024 (Dec. 10-15) to present our papers on causal representation learning, causal discovery, and causal bandits (see thread). Looking to meet old and new friends and chat on causality / representation learning.
0
0
2
Great reminder how our thinking is (often subconsciously & heavily) influenced by the precedence and being generous with acknowledging that is important.
@mikeroberts3000 @docmilanfar Schmidhuber was working in the same field on the same artefacts just two decades too early. So I don’t interpret Schmidhuber as shit posting often. My post is about the statement “Attention was not discovered in machine learning”. This is a bad way to think about discovery, in.
0
0
3
The P&T letter I was working on was due last night. Now the delay is partly on you @overleaf!.
⚠️ Attention: our website is down. We are aware of the outage and are investigating. Apologies for any inconvenience this might cause. You can also check our status page at
0
0
1
I’ll be more active on Bluesky and look forward to (re)connecting there:.
0
0
1
RT @rpi: President Marty Schmidt ’81 spoke at @economistimpact's 3rd annual Commercialising Quantum Global event this week. “What does bein….
0
5
0