Peter Vamplew
amp1874.bsky.social
Peter Vamplew
@amp1874.bsky.social
Professor in IT @ Federation Uni. Multi-objective reinforcement learning. Human-aligned AI. Best known for the f*cking mailing list paper. Jambo & Bengals fan. https://t.co/UNoOrbGApz
Pinned
I'm going to post about a few of my key and/or recent papers to create some context for my bsky profile.

Probably the most important paper I've been part of: link.springer.com/article/10.1...
This practical guide reviews the 'why' and 'how' of multi-objective reinforcement learning.
A practical guide to multi-objective reinforcement learning and planning - Autonomous Agents and Multi-Agent Systems
Real-world sequential decision-making tasks are generally complex, requiring trade-offs between multiple, often conflicting, objectives. Despite this, the majority of research in reinforcement learnin...
link.springer.com
Reposted by Peter Vamplew
Getting nervous for the talk I'm about to give at a workshop about "using AI to drive impact" which features slides such as these.
November 6, 2025 at 8:41 PM
I guess I should feel flattered that AI models keep hallucinating papers that I didn't actually write?

The latest example is "Where reinforcement learning meets process control" which I apparently co-authored with Kumar and Yu. #ADIPEC 1/n
November 6, 2025 at 10:01 AM
Reposted by Peter Vamplew
@clarivate.com says it takes retractions into account when calculating its highly coveted researcher designations. But Qusay Hassan, who has had 21 papers retracted, was one of several scientists winning accolades at the ministry’s Iraq Education Conference 2025 in Baghdad earlier this month.
Exclusive: Web of Science company involved in dubious awards in Iraq
Hayder A. Dhahad, Iraq’s deputy minister for scientific research affairs, speaks at an awards ceremony at the country’s Science Day celebration. Source: Instagram In the string of prestigious…
retractionwatch.com
October 31, 2025 at 10:08 AM
Reposted by Peter Vamplew
New #J2C Certification:

Demonstration-Guided Multi-Objective Reinforcement Learning

Junlin Lu, Patrick Mannion, Karl Mason

https://openreview.net/forum?id=FQAgFgkaFG

#reinforcement #demonstrations #objective
October 26, 2025 at 12:23 AM
I'm feeling much more organised after tidying up my office yesterday. I finally have all my ducks in a row.
October 26, 2025 at 1:04 AM
Congratulations to @ffelten.bsky.social whose thesis on multi-objective reinforcement learning was deservedly awarded the Luxembourg National Research Fund's Outstanding PhD Thesis Award 🏆

As a side-benefit, we get this great accessible introduction to MORL video: youtu.be/VEXRuhJDkoA
FNR Awards 2025: Outstanding PhD Thesis - Florian Felten
YouTube video by FNRLux
youtu.be
October 26, 2025 at 12:59 AM
It's been 8 months since I notified @springernature.com of a clear-cut case of fake references in a book chapter which they published. They have still not taken any action as they claim to still be investigating.

I will not review for Springer again until this matter is satisfactorily resolved.
October 21, 2025 at 11:27 AM
Hey @springernature.com. What was the point in me setting my status to Paused on the Reviewer Dashboard yesterday if you're going to send me another review today anyway? Not to mention that paper was 99% unrelated to my area of research. Don't turn into MDPI.
October 20, 2025 at 9:04 PM
Reposted by Peter Vamplew
The viral "Definition of AGI" paper tells you to read fake references which do not exist!

Proof: different articles present at the specified journal/volume/page number, and their titles exist nowhere on any searchable repository.

Take this as a warning to not use LMs to generate your references!
October 18, 2025 at 12:54 AM
Is anyone else on Overleaf's free plan? Have you had issues with compile time-outs? I gather that the time-limit for compilation on the free plan was cut in Aug, but it's only in the last few days that I've had problems. Today I can't even compile a 2-page document with no tables or figures.
October 16, 2025 at 1:53 AM
Dear Benjamin,

Congrats on being the fastest scammy conference organiser of all time. Inviting me to a conference unrelated to the topic of my paper is highly questionable, but you did it within a day of publication so at least you’re fast.

Kindly remove me from your mailing list.

Regards,
Peter
October 8, 2025 at 1:30 AM
Computer says sorry?

After months in copy-editing hell, Haddie Harland's review of AI apology research is now available: link.springer.com/article/10.1...

This is a must read for anyone interested in how AI systems can effectively and appropriately use apologies to facilitate human interaction 1/2
AI apology: a critical review of apology in AI systems - Artificial Intelligence Review
Apologies are a powerful tool used in human-human interactions to provide affective support, regulate social processes, and exchange information following a trust violation. The emerging field of AI apology investigates the use of apologies by artificially intelligent systems, with recent research suggesting how this tool may provide similar value in human-machine interactions. Until recently, contributions to this area were sparse, and these works have yet to be synthesised into a cohesive body of knowledge. This article provides the first synthesis and critical analysis of the state of AI apology research, focusing on studies published between 2020 and 2023. We derive a framework of attributes to describe five core elements of apology: outcome, interaction, offence, recipient, and offender. With this framework as the basis for our critique, we show how apologies can be used to recover from misalignment in human-AI interactions, and examine trends and inconsistencies within the field. Among the observations, we outline the importance of curating a human-aligned and cross-disciplinary perspective in this research, with consideration for improved system capabilities and long-term outcomes.
link.springer.com
October 6, 2025 at 10:13 PM
Decisions, decisions. Which should I read first?
October 3, 2025 at 3:05 AM
Reposted by Peter Vamplew
Dsouza, Ofosu, Amaogu, Pigeon, Boudreault, Maghoul, Moreno-Cruz, Leonenko: BoreaRL: A Multi-Objective Reinforcement Learning Environment for Climate-Adaptive Boreal Forest Management https://arxiv.org/abs/2509.19846 https://arxiv.org/pdf/2509.19846 https://arxiv.org/html/2509.19846
September 25, 2025 at 6:33 AM
Reposted by Peter Vamplew
Lingxiao Kong, Cong Yang, Oya Deniz Beyan, Zeyd Boukhers
Multi-Objective Reinforcement Learning for Large Language Model Optimization: Visionary Perspective
https://arxiv.org/abs/2509.21613
September 29, 2025 at 10:02 AM
Reposted by Peter Vamplew
The deadline for my postdoc on scalable clinical decision support is closing in 1 week: 4 October (Australian Eastern standard Time). Please share with anyone that you think would be interested
I'm hiring again! Please share. I'm recruiting a postdoc research fellow in human-centred AI for scalable decision support. Join us to investigate how to balance scalability and human control in medical decision support. Closing date: 4 October (AEST).
uqtmiller.github.io/recruitment/
Recruitment
uqtmiller.github.io
September 26, 2025 at 12:42 AM
Dear authors who I shall not name. Thank you for citing my work. But I'm not sure that a paper dating from 1995 should be cited in the context of a paragraph which begins "Recent trends show...."
September 24, 2025 at 7:12 AM
Reposted by Peter Vamplew
⚠️ The #CHI2026 paper I submitted? It almost didn't exist. That's the BTS part academics never post. So I will…to normalize what I call unglamorous persistence.

This summer was one of my hardest, mentally. 🌥️ Between ...
1/n
September 22, 2025 at 4:51 PM
Hey! There's finally someone else in Australia doing research in multi-objective reinforcement learning. @marcusgal.bsky.social arxiv.org/pdf/2509.14816
arxiv.org
September 21, 2025 at 12:44 PM
Reposted by Peter Vamplew
September 17, 2025 at 2:35 AM
Almost 7 months ago I reported a book chapter to Springer's Ethics reporting team. Springer have yet to take any action (they are 'still investigating'). That chapter now has 450 downloads and 2 citations. Delays in dealing with these issues allow further pollution of the scientific literature.
September 17, 2025 at 12:50 AM
Reposted by Peter Vamplew
I'm hiring again! Please share. I'm recruiting a postdoc research fellow in human-centred AI for scalable decision support. Join us to investigate how to balance scalability and human control in medical decision support. Closing date: 4 October (AEST).
uqtmiller.github.io/recruitment/
Recruitment
uqtmiller.github.io
September 16, 2025 at 4:34 AM
Reposted by Peter Vamplew
AI-SearchPlanner: Modular Agentic Search via Pareto-Optimal Multi-Objective Reinforcement Learning

Huawei proposes an RL framework that decouples search planning from answer generation, using dual-reward alignment and Pareto optimization.

📝 arxiv.org/abs/2508.20368
AI-SearchPlanner: Modular Agentic Search via Pareto-Optimal Multi-Objective Reinforcement Learning
Recent studies have explored integrating Large Language Models (LLMs) with search engines to leverage both the LLMs' internal pre-trained knowledge and external information. Specially, reinforcement l...
arxiv.org
August 29, 2025 at 2:29 AM
Reposted by Peter Vamplew
Multi-objective reinforcement learning (MORL) offers a more robust and adaptable solution by optimizing for a vector of rewards—such as fairness, diversity, and ethical norms. The resulting behaviors support transparency, explainability, and human alignment.
September 4, 2025 at 1:14 PM