We study a $K$-armed non-stationary bandit model where rewards change smoothly, as captured by H\"{o}lder class assumptions on rewards as functions of time. Such smooth changes are parametrized by a H\"{o}lder exponent $\beta$ and coefficient $\lambda$. While various sub-cases of this general model have been studied in isolation, we first establish the minimax dynamic regret rate generally for all $K,\beta,\lambda$. Next, we show this optimal dynamic regret can be attained adaptively, without knowledge of $\beta,\lambda$. To contrast, even with parameter knowledge, upper bounds were only previously known for limited regimes $\beta\leq 1$ and $\beta=2$ (Slivkins, 2014; Krishnamurthy and Gopalan, 2021; Manegueu et al., 2021; Jia et al.,2023). Thus, our work resolves open questions raised by these disparate threads of the literature. We also study the problem of attaining faster gap-dependent regret rates in non-stationary bandits. While such rates are long known to be impossible in general (Garivier and Moulines, 2011), we show that environments admitting a safe arm (Suk and Kpotufe, 2022) allow for much faster rates than the worst-case scaling with $\sqrt{T}$. While previous works in this direction focused on attaining the usual logarithmic regret bounds, as summed over stationary periods, our new gap-dependent rates reveal new optimistic regimes of non-stationarity where even the logarithmic bounds are pessimistic. We show our new gap-dependent rate is tight and that its achievability (i.e., as made possible by a safe arm) has a surprisingly simple and clean characterization within the smooth H\"{o}lder class model.

Were you unable to attend Transform 2022? Check out all of the summit sessions in our on-demand library now! Making artificial intelligence (AI) work with humans requires having an internal data champion to help overcome fears and create a safe environment. That was the advice offered at Digital Procurement World in Amsterdam today in a session focused on combining humans and AI to help procurement become a top value driver in modern business. Working through overwhelming amounts of data and dealing with employee concerns that AI will take their jobs.

Satisfying safety constraints almost surely (or with probability one) can be critical for deployment of Reinforcement Learning (RL) in real-life applications. For example, plane landing and take-off should ideally occur with probability one. We address the problem by introducing Safety Augmented (Saute) Markov Decision Processes (MDPs), where the safety constraints are eliminated by augmenting them into the state-space and reshaping the objective. We show that Saute MDP satisfies the Bellman equation and moves us closer to solving Safe RL with constraints satisfied almost surely. We argue that Saute MDP allows to view Safe RL problem from a different perspective enabling new features.

In the last few years, we've been experimenting with humanlike, Artificial Intelligence-based avatars. We integrated these avatars into L&D environments, workplaces, and the academic field. These avatars can act both as personal mentors (or trainers) and as clients or workers in real-world simulations. We wanted to share some of our insights. We believe AI-based trainers are going to dramatically change corporate learning, and empower both trainers and learners.