Related papers: Comment on "Learning to Build the Bomb"
AI models need to be unlearned to fulfill the requirements of legal acts such as the AI Act or GDPR, and also because of the need to remove toxic content, debiasing, the impact of malicious instances, or changes in the data distribution…
Current content filtering and blocking methods are susceptible to various circumvention techniques and are relatively slow in dealing with new threats. This is due to these methods using shallow pattern recognition that is based on regular…
This paper investigates a problem about freedom of information. Although freedom of information is generally considered desirable, there are a number of areas where there is substantial agreement that freedom of information should be…
After more than 80 years from the seminal work of Weizs\"acker and the liquid drop model of the atomic nucleus, deviations from experiments of mass models ($\sim$ MeV) are orders of magnitude larger than experimental errors ($\lesssim$…
Understanding of the atomic structures and ways in which the atoms interacting is critical to the understanding of chemistry. In this paper applied the concepts of game theory in explaining reactions between elements of the periodic table.…
The use of Reinforcement Learning (RL) is still restricted to simulation or to enhance human-operated systems through recommendations. Real-world environments (e.g. industrial robots or power grids) are generally designed with safety…
Selective classification allows models to abstain from making predictions (e.g., say "I don't know") when in doubt in order to obtain better effective accuracy. While typical selective models can be effective at producing more accurate…
The article recently published by M. Gai et al. claims to reach conclusions from a collaborative experiment dedicated to the study of the 7Be(n,alpha) reaction. These claims were published with no authorization from key collaborators,…
Information sharing is vital in resisting cyberattacks, and the volume and severity of these attacks is increasing very rapidly. Therefore responders must triage incoming warnings in deciding how to act. This study asked a very specific…
Enrico Fermi and Leo Szilard worked together at Columbia in 1939-40, just after nuclear fission was discovered, to ascertain the feasibility of a nuclear chain reaction, and then on the construction of the first nuclear reactor. Szilard…
How does the extent to which a model is open or closed impact the scientific inferences that can be drawn from research that involves it? In this paper, we analyze how restrictions on information about model construction and deployment…
Cooperative dynamics are central to our understanding of many phenomena in living and complex systems. However, we lack a universal mechanism to explain the emergence of cooperation. We present a novel framework for modelling social dilemma…
This report is based on a series of papers written between 1980 and 2005 on the origin of Iraq's nuclear weapons program, which was known to one of the authors in the late 1970s already, as well as to a number of other physicists, who…
I offer an analysis of the Principle of Sufficient Reason and its relevancy for the scientific endeavour. I submit that the world is not, and cannot be, rational - only some brained beings are. The Principle of Sufficient Reason is not a…
In his Comment on our recent paper ``The laws of physics do not prohibit counterfactual communication'', \textit{npj Quantum Information} (2022) 8:60, Popescu argues that the claims of the paper are invalid. Here, we refute his argument,…
The rapid expansion of scientific literature makes it increasingly difficult to acquire new knowledge, particularly in specialized domains where reasoning is complex, full-text access is restricted, and target references are sparse among a…
Existing question-answering research focuses on unanswerable questions in the context of always providing an answer when a system can\dots but what about cases where a system {\bf should not} answer a question. This can either be to protect…
Safe reinforcement learning is a promising path toward applying reinforcement learning algorithms to real-world problems, where suboptimal behaviors may lead to actual negative consequences. In this work, we focus on the setting where…
This paper offers a new perspective on the limits of machine learning: the ceiling on progress is set not by model size or algorithm choice but by the information structure of the task itself. Code generation has progressed more reliably…
International verification of nuclear warheads is a practical problem in which the protection of secret warhead information is of paramount importance. We propose a measure that would enable a weapon owner to evaluate the privacy of a…