Sign in

Murat Kocaoglu

@murat-kocaoglu.bsky.social
150 followers 102 following 35 posts

Asst. Prof. at Purdue ECE. Causal ML Lab. Causal discovery, causal inference, deep generative models, info theory, online learning. Past: MIT-IBM AI Lab, UT Austin, Koc, METU.

PostsRepliesMedia
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/02/2026
AISTATS will take place in beautiful Tangier, Morocco on May 5th, 2026. Your 4-page pdf may contain either new results or content from a relevant paper of yours published since 2025. We have a fantastic lineup of speakers and panelists!
010
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/02/2026
I am excited to announce our Workshop on Causality in the Age of AI Scaling in AISTATS 2026! - Is scaling sufficient for intelligent systems? - Can causal abilities emerge from scale? - What can causal modeling bring that scale cannot? causcale.github.io Submission deadline is February 27th.
causcale.github.io
Causality in the Age of AI Scaling
Workshop on Causality in the Age of AI Scaling (5 May 2026, Tangier, Morocco)
171
Murat Kocaoglu @murat-kocaoglu.bsky.social · 31/10/2025
Happy to be a part of this cohort! Apply to the JHU CS PhD program to work together. engineering.jhu.edu/admissions/g...
t.co
https://engineering.jhu.edu/admissions/graduate-admissions/full-time-programs/how-to-apply/
030
Reposted by Murat Kocaoglu
JHU Computer Science @jhucompsci.bsky.social · 08/10/2025
10 new CS professors! 🥳 @anandbhattad.bsky.social @uthsav.bsky.social @gligoric.bsky.social @murat-kocaoglu.bsky.social @tiziano.bsky.social
0106
Murat Kocaoglu @murat-kocaoglu.bsky.social · 05/06/2025
Thank you so much Naftali! Hope all is well
000
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/06/2025
Thank you!
010
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/06/2025
Bookmark our lab page and GitHub repo to follow our work: muratkocaoglu.com/CausalML/ github.com/CausalML-Lab
muratkocaoglu.com
CausalML Lab
000
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/06/2025
CausalML Lab will continue to push the boundaries of fundamental causal inference and discovery research with an added focus on real-world applications and impact. If you are at Johns Hopkins @jhu.edu, or more generally on the East Coast, and are interested in collaborating, please reach out.
110
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/06/2025
I am also deeply grateful to Purdue University and @purdueece.bsky.social for their support during my first four years as a professor. I had the privilege of teaching enthusiastic undergrads, working with outstanding PhD students, and great colleagues there. I learned a great deal from them.
111
Murat Kocaoglu @murat-kocaoglu.bsky.social · 03/06/2025
I am happy to share that I will be joining Johns Hopkins University's Computer Science Department @jhucompsci.bsky.social as an Assistant Professor in Fall 2025. I am grateful to my mentors for their unwavering support and to my exceptional PhD students for advancing our lab's research vision.
270
Murat Kocaoglu @murat-kocaoglu.bsky.social · 25/01/2025
Faithfulness can be relaxed significantly. Determinism (?) not sure what this means but I don't think it's equivalent to CMC. No unmeasured confounder is not necessary. You need some sparsity to observe some independence pattern, that's all; to learn something, but not necessarily everything.
020
Murat Kocaoglu @murat-kocaoglu.bsky.social · 28/12/2024
They used to give champagne glass? I was surprised to see something other than a mug this year.
120
Murat Kocaoglu @murat-kocaoglu.bsky.social · 24/12/2024
Teaching young researchers mathematical and scientific rigor is more important than ever today with AI tools' wider adoption in research, as these tools tend to be overly optimistic. LLM-assisted false proofs risk flooding our already overloaded reviewing infrastructure.
041
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
We will present this work at #NeurIPS2024 on Wednesday at 4:30pm local time in Vancouver. Poster #5107. Led by my PhD students Zihan Zhou and Qasim Elahi. Paper link: openreview.net/forum?id=RfS... Follow us for more updates from the #CausalML Lab!
openreview.net
Sample Efficient Bayesian Learning of Causal Graphs from Interventions
Causal discovery is a fundamental problem with applications spanning various areas in science and engineering. It is well understood that solely using observational data, one can only orient the...
010
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
Wienöbst et al.'s way of uniformly sampling from Markov equivalent DAGs allows us to answer other interesting questions. We focus on estimating the causal effect of non-manipulable variables. We can learn the edges adjacent to this node (a graph cut) and use adjustment from obs.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
We then update the posteriors over each graph cut, which quickly converge to the true cut configurations. This gives us a sample-efficient way to learn causal graphs through interventions non-parametrically for discrete variables. Green at the bottom is ours vs. some baselines.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
Assuming we have enough obs. data, we can compute the likelihood of our intv. samples given any graph cut, even though we don't know the graph. Two ways to do this are by unif. sampling causal graphs in poly-time thanks to Wienöbst et al. 2023, or by adjustment of Perkovic 2020.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
We leverage an idea from 2010s on learning causal graphs with small interventions. (n, k) separating systems cut every edge with interventions of size k. Each intervention gives info about a cut. We keep track of the posteriors of a set of graph cuts due to (n, k) sep. system.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
Bayesian approaches are promising since they can incorporate causal knowledge even in a single interventional sample. However they are computationally intensive to run on large graphs. Instead of keeping track of all causal DAGs, can we keep track of a compact set of subgraphs?
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 10/12/2024
You have causal questions. First step is deciding cause-effect relations in your system. Causal graphs capture these compactly. We usually need to be able to run experiments and use the outcomes to infer causal graphs. But experiments are expensive & w/ few samples. #NeurIPS2024
110
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
I will present this work at #NeurIPS2024 next Thursday at 11am local time in Vancouver. Poster #5104. Led by my PhD student Qasim Elahi. Joint work with my colleague Mahsa Ghasemi. Paper link: openreview.net/forum?id=uM3... Follow us for more updates from the #CausalML Lab!
openreview.net
Partial Structure Discovery is Sufficient for No-regret Learning in...
Causal knowledge about the relationships among decision variables and a reward variable in a bandit setting can accelerate the learning of an optimal decision. Current works often assume the causal...
011
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
Finally, we have our bandit algorithm that can operate in unknown environments taking advantage of the fact that partial causal discovery is sufficient for achieving optimal regret, pseudocode below:
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
A toy example from the paper: Missing V1 <--> V3 does not affect the possibly optimal minimal intervention sets (POMIS), missing any other bidirected edge does. So we don't need to allocate rounds in our causal bandit algorithm for learning this edge after learning the rest.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
We find that not all confounder locations are needed. You can get away with not learning some and you end up with the same POMIS set, which means you will never miss an optimal arm! We propose an interventional causal discovery algorithm that takes advantage of this observation.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
If you don't know the causal graph, you may try to learn it from data and/or experiments. We identify an interesting research question here: Do we need to know all unobserved confounders to learn all POMISes? Or can we get away without knowing some? This is not obvious.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
A very nice idea is the POMIS developed by Sanghack Lee and Elias Bareinboim in 2018. They use do-calculus to eliminate unnecessary actions that give the same reward. They propose a principled algorithm to do this. And if you use less, they show you may miss the optimal arm.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
Most existing work focuses on how this action space reduction can be done algorithmically if you know the causal structure. In a semi-Markovian model, this includes the location of every unobserved confounder represented as bidirected edge.
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
You want to optimize a reward in an unknown environment. Structural knowledge of cause-effect relations is known to help significantly reduce the search space for bandit algorithms. But how much of the causal structure do you need to know to do this? #NeurIPS2024
121
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
Not exactly sure what that means or how to do that. Any link with more info on this?
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
We will present this work at #NeurIPS2024 next Thursday at 11am local time in Vancouver. Poster #5103. Joint work led by my PhD student Md. Musfiqur Rahman and colleague Matt Jordan. Paper link: openreview.net/forum?id=vym... Follow us for more updates from the #CausalML Lab!
openreview.net
Conditional Generative Models are Sufficient to Sample from Any...
Causal inference from observational data plays critical role in many applications in trustworthy machine learning. While sound and complete algorithms exist to compute causal effects, many of them...
010
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
With our method, we can quantify how spurious correlations in their training data affect large image generative models. For example, we can quantify how much changing the biological sex of a person affects their perceived age, a non-causal relation that shouldn't be there:
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
Our algorithm mimics the steps of the ID algorithm and inherits its soundness and completeness guarantees for arbitrary causal queries. It can also seamlessly sample from conditional interventional queries; the existing deep causal generative models use rejection sampling (slow)
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
Our paper answers this question positively. We propose an algorithm that trains a bunch of conditional generators, and ties them together to sample from any causal effect estimand. This unlocks the potential of diffusion models for causal inference!
220
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
Deep generative models give us a way to sample from such conditional distributions, learning them implicitly. A fundamental question we ask is, can we sample from any identifiable causal query (distribution) just by using conditional generative models?
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
With enough data, we know which causal questions can be answered and which cannot thanks to the ID algorithms of Tian & Pearl and Shpitser & Pearl. But these require the likelihood of complicated high dimensional distributions. These can't be explicitly learned from data, e.g.,
100
Murat Kocaoglu @murat-kocaoglu.bsky.social · 08/12/2024
You have a complicated system with several high-dimensional variables, such as image and text data. How can you systematically answer ANY causal question from observational data? #NeurIPS2024
261