Marlos C. Machado
@MarlosCMachado
Assistant Professor @UAlberta. @AmiiThinks Fellow, Canada CIFAR AI chair.
Now that @RL_Conference is over (and what a success it was!), it's time to start thinking about the next one. If you haven't heard yet, RLC is going north. We'll be organizing RLC'25 in Edmonton! I'm very excited about that and look forward to seeing everyone there.
Back in April, Amii and @UAlberta launched Vulcan, a high-performance compute site that is set to advance AI and the discovery of novel solutions in Alberta. The initiative is a key part of the Pan-Canadian AI Compute Environment project. Read: hubs.ly/Q03yYJCG0
My partner created an Instagram account to talk about her work as a registered psychologist focused on trauma and perinatal mental health. If this is something that interests you, here's the link: instagram.com/_tidesofbecomi…
Previous work has shown that programmatic policies—computer programs written in a domain-specific language—generalize to out-of-distribution problems more easily than neural policies. Is this really the case? 🧵
Prabhat's work is an extremely careful empirical investigation of double learning in deep RL. I think this is a great paper that people should be aware of if they care about deep RL.
(1/7) We (me, @white_martha, @MarlosCMachado) have a new preprint: "Double Q-learning for Value-based Deep Reinforcement Learning, Revisited", answering a question that has been lingering for a decade! Stick around for the arXiv link.
This was a great long-term effort from @MartinKlissarov, @akhil_bagaria, and @RayZiyan41307, and it led to a great overview of the ideas behind leveraging temporal abstractions in AI. If anything, I think this is very useful resource for anyone interested in this field!
As AI agents face increasingly long and complex tasks, decomposing them into subtasks becomes increasingly appealing. But how do we discover such temporal structure? Hierarchical RL provides a natural formalism-yet many questions remain open. Here's our overview of the field🧵
I second that.
I like em dashes, and I hate that some people now take them as evidence of text being AI generated.
A new opening for multimodal model research: jobs.careers.microsoft.com/global/en/job/… . Please apply if interested.
🧵 Solving needle-in-a-haystack problems in planning is hard. The search attempts and fails to solve such problems many times before finding a solution it can learn from. @JakeTuero's method learns subgoals from failed searches to train policies. arxiv.org/pdf/2506.07255
Phaidra is hiring a Research Scientist to work on sequential decision-making problems. I'm at the RLDM conference in Dublin this week. If you're attending and would like to learn more about the role or the company, feel free to reach out! job-boards.greenhouse.io/phaidra/jobs/4…
I will be presenting a poster at RLDM on Wednesday @ 4:30pm. We show that embedding an agent implicitly constrains both it and the environment. We use this constraint to characterize continual adaptation. lewandowskialex.com/papers/files/l… Feel free to reach out if you're attending RLDM!
Excited to share that I'll be joining the CS department at @UAlberta as an Assistant Professor in January 2026 where I will be affiliated with @AmiiThinks. I'll be recruiting 2-3 PhD/MSc students and establishing a research lab on AI in Education and Human–AI Interaction.
This week, we have the honor of hosting @MarlosCMachado as our guest speaker.
Publicizing this: if you have an accepted or published paper in IEEE Transactions on Automatic Control IEEE Transactions on Control of Network Systems IEEE Open Journal on Control Systems *without a conference version*, you can apply to present it at CDC 2025 in Brazil! All…
To align better with workshop acceptance dates, 𝐑𝐋𝐂 𝐢𝐬 𝐞𝐱𝐭𝐞𝐧𝐝𝐢𝐧𝐠 𝐢𝐭𝐬 𝐞𝐚𝐫𝐥𝐲 𝐫𝐞𝐠𝐢𝐬𝐭𝐫𝐚𝐭𝐢𝐨𝐧 𝐝𝐞𝐚𝐝𝐥𝐢𝐧𝐞 𝐭𝐨 𝐉𝐮𝐧𝐞 𝟐𝟑𝐫𝐝!
Deadlines extended. Programmatic Representations for Agent Learning (ICML) - Vancouver. - Deadline: May 30, 2025 (pral-workshop.github.io) Programmatic Reinforcement Learning (RLC) - Edmonton. - Deadline: June 6, 2025 (prl-workshop.github.io)
Are you interested in programmatic representations in machine learning? Please consider submitting your work to these workshops: one at ICML and another at RLC.