In imperfect information games (e.g. Bridge, Skat, Poker), one of the fundamental considerations is to infer the missing information while at the same time avoiding the disclosure of private information. Disregarding the issue of protecting private information can lead to a highly exploitable performance. Yet, excessive attention to it leads to hesitations that are no longer consistent with our private information. In our work, we show that to improve performance, one must choose whether to use a player's private information. We extend our work by proposing a new belief distribution depending on the amount of private and public information desired. We empirically demonstrate an increase in performance and, with the aim of further improving performance, the new distribution should be used according to the position in the game. Our experiments have been done on multiple benchmarks and in multiple determinization-based algorithms (PIMC and IS-MCTS).

Autonomous navigation in environments is a critical capability for modern mobile robots, and has been extensively studied over several decades. Classical approaches to navigation rely on constructing detailed maps of the environment and accurate localization of the robot within the map [1, 2, 3]. However, with increasing demand for deploying robots in novel uncontrolled environments such as households, last-mile delivery, etc., constructing accurate and fine-grained maps frequently is often impractical. Robots must now be able to navigate without maps, which means efficient navigation policies require accurate semantic understanding of the scene, efficient exploration and episodic memory, and long-horizon planning with limited knowledge of the environment. Advances in scene understanding and a have led to semantic navigation tasks such as image-goal navigation [4, 5], object-goal navigation [6, 7], etc. receiving significant focus in recent years. In this work, we consider the specific task of image-goal navigation where robot's navigation objective is specified by an RGB image. We motivate the task with the following scenario: Consider a mobile household robot equipped with an onboard camera tasked with picking up a novel unseen object (say, a new shirt). Since the robot has no prior knowledge about the novel object, it would need other semantic information to understand the object - an image of the object would serve this purpose effectively.

Predicting cancer dynamics under treatment is challenging due to high inter-patient heterogeneity, lack of predictive biomarkers, and sparse and noisy longitudinal data. Mathematical models can summarize cancer dynamics by a few interpretable parameters per patient. Machine learning methods can then be trained to predict the model parameters from baseline covariates, but do not account for uncertainty in the parameter estimates. Instead, hierarchical Bayesian modeling can model the relationship between baseline covariates to longitudinal measurements via mechanistic parameters while accounting for uncertainty in every part of the model. The mapping from baseline covariates to model parameters can be modeled in several ways. A linear mapping simplifies inference but fails to capture nonlinear covariate effects and scale poorly for interaction modeling when the number of covariates is large. In contrast, Bayesian neural networks can potentially discover interactions between covariates automatically, but at a substantial cost in computational complexity. In this work, we develop a hierarchical Bayesian model of subpopulation dynamics that uses baseline covariate information to predict cancer dynamics under treatment, inspired by cancer dynamics in multiple myeloma (MM), where serum M protein is a well-known proxy of tumor burden. As a working example, we apply the model to a simulated dataset and compare its ability to predict M protein trajectories to a model with linear covariate effects. Our results show that the Bayesian neural network covariate effect model predicts cancer dynamics more accurately than a linear covariate effect model when covariate interactions are present. The framework can also be applied to other types of cancer or other time series prediction problems that can be described with a parametric model.

Predicting simple function classes has been widely used as a testbed for developing theory and understanding of the trained Transformer's in-context learning (ICL) ability. In this paper, we revisit the training of Transformers on linear regression tasks, and different from all the existing literature, we consider a bi-objective prediction task of predicting both the conditional expectation $\mathbb{E}[Y|X]$ and the conditional variance Var$(Y|X)$. This additional uncertainty quantification objective provides a handle to (i) better design out-of-distribution experiments to distinguish ICL from in-weight learning (IWL) and (ii) make a better separation between the algorithms with and without using the prior information of the training distribution. Theoretically, we show that the trained Transformer reaches near Bayes-optimum, suggesting the usage of the information of the training distribution. Our method can be extended to other cases. Specifically, with the Transformer's context window $S$, we prove a generalization bound of $\tilde{\mathcal{O}}(\sqrt{\min\{S, T\}/(n T)})$ on $n$ tasks with sequences of length $T$, providing sharper analysis compared to previous results of $\tilde{\mathcal{O}}(\sqrt{1/n})$. Empirically, we illustrate that while the trained Transformer behaves as the Bayes-optimal solution as a natural consequence of supervised training in distribution, it does not necessarily perform a Bayesian inference when facing task shifts, in contrast to the \textit{equivalence} between these two proposed in many existing literature. We also demonstrate the trained Transformer's ICL ability over covariates shift and prompt-length shift and interpret them as a generalization over a meta distribution.

The process of calibrating computer models of natural phenomena is essential for applications in the physical sciences, where plenty of domain knowledge can be embedded into simulations and then calibrated against real observations. Current machine learning approaches, however, mostly rely on rerunning simulations over a fixed set of designs available in the observed data, potentially neglecting informative correlations across the design space and requiring a large amount of simulations. Instead, we consider the calibration process from the perspective of Bayesian adaptive experimental design and propose a data-efficient algorithm to run maximally informative simulations within a batch-sequential process. At each round, the algorithm jointly estimates the parameters of the posterior distribution and optimal designs by maximising a variational lower bound of the expected information gain. The simulator is modelled as a sample from a Gaussian process, which allows us to correlate simulations and observed data with the unknown calibration parameters. We show the benefits of our method when compared to related approaches across synthetic and real-data problems.

Bayesian Optimization (BO) has proven to be very successful at optimizing a static, noisy, costly-to-evaluate black-box function $f : \mathcal{S} \to \mathbb{R}$. However, optimizing a black-box which is also a function of time (i.e., a dynamic function) $f : \mathcal{S} \times \mathcal{T} \to \mathbb{R}$ remains a challenge, since a dynamic Bayesian Optimization (DBO) algorithm has to keep track of the optimum over time. This changes the nature of the optimization problem in at least three aspects: (i) querying an arbitrary point in $\mathcal{S} \times \mathcal{T}$ is impossible, (ii) past observations become less and less relevant for keeping track of the optimum as time goes by and (iii) the DBO algorithm must have a high sampling frequency so it can collect enough relevant observations to keep track of the optimum through time. In this paper, we design a Wasserstein distance-based criterion able to quantify the relevancy of an observation with respect to future predictions. Then, we leverage this criterion to build W-DBO, a DBO algorithm able to remove irrelevant observations from its dataset on the fly, thus maintaining simultaneously a good predictive performance and a high sampling frequency, even in continuous-time optimization tasks with unknown horizon. Numerical experiments establish the superiority of W-DBO, which outperforms state-of-the-art methods by a comfortable margin.

Multi-dimensional parameter spaces are commonly encountered in astroparticle physics theories that attempt to capture novel phenomena. However, they often possess complicated posterior geometries that are expensive to traverse using techniques traditional to this community. Effectively sampling these spaces is crucial to bridge the gap between experiment and theory. Several recent innovations, which are only beginning to make their way into this field, have made navigating such complex posteriors possible. These include GPU acceleration, automatic differentiation, and neural-network-guided reparameterization. We apply these advancements to astroparticle physics experimental results in the context of novel neutrino physics and benchmark their performances against traditional nested sampling techniques. Compared to nested sampling alone, we find that these techniques increase performance for both nested sampling and Hamiltonian Monte Carlo, accelerating inference by factors of $\sim 100$ and $\sim 60$, respectively. As nested sampling also evaluates the Bayesian evidence, these advancements can be exploited to improve model comparison performance while retaining compatibility with existing implementations that are widely used in the natural sciences.

We resolve the open problem of designing a computationally efficient algorithm for infinite-horizon average-reward linear Markov Decision Processes (MDPs) with $\widetilde{O}(\sqrt{T})$ regret. Previous approaches with $\widetilde{O}(\sqrt{T})$ regret either suffer from computational inefficiency or require strong assumptions on dynamics, such as ergodicity. In this paper, we approximate the average-reward setting by the discounted setting and show that running an optimistic value iteration-based algorithm for learning the discounted setting achieves $\widetilde{O}(\sqrt{T})$ regret when the discounting factor $\gamma$ is tuned appropriately. The challenge in the approximation approach is to get a regret bound with a sharp dependency on the effective horizon $1 / (1 - \gamma)$. We use a computationally efficient clipping operator that constrains the span of the optimistic state value function estimate to achieve a sharp regret bound in terms of the effective horizon, which leads to $\widetilde{O}(\sqrt{T})$ regret.

Continuous normalizing flows (CNFs) learn the probability path between a reference and a target density by modeling the vector field generating said path using neural networks. Recently, Lipman et al. (2022) introduced a simple and inexpensive method for training CNFs in generative modeling, termed flow matching (FM). In this paper, we re-purpose this method for probabilistic inference by incorporating Markovian sampling methods in evaluating the FM objective and using the learned probability path to improve Monte Carlo sampling. We propose a sequential method, which uses samples from a Markov chain to fix the probability path defining the FM objective. We augment this scheme with an adaptive tempering mechanism that allows the discovery of multiple modes in the target. Under mild assumptions, we establish convergence to a local optimum of the FM objective, discuss improvements in the convergence rate, and illustrate our methods on synthetic and real-world examples.

Fairness is a critical objective in policy design and algorithmic decision-making. Identifying the causal pathways of unfairness requires knowledge of the underlying structural causal model, which may be incomplete or unavailable. This limits the practicality of causal fairness analysis in complex or low-knowledge domains. To mitigate this practicality gap, we advocate for developing efficient causal discovery methods for fairness applications. To this end, we introduce local discovery for direct discrimination (LD3): a polynomial-time algorithm that recovers structural evidence of direct discrimination. LD3 performs a linear number of conditional independence tests with respect to variable set size. Moreover, we propose a graphical criterion for identifying the weighted controlled direct effect (CDE), a qualitative measure of direct discrimination. We prove that this criterion is satisfied by the knowledge returned by LD3, increasing the accessibility of the weighted CDE as a causal fairness measure. Taking liver transplant allocation as a case study, we highlight the potential impact of LD3 for modeling fairness in complex decision systems. Results on real-world data demonstrate more plausible causal relations than baselines, which took 197x to 5870x longer to execute.