This work defines a complexity measure for BAMDP planning and introduces state abstraction for more efficient approximate planning.
problem The computational intractability of exact BAMDP planning solutions.
method Define a complexity measure for BAMDP planning, introduce state abstraction, and develop an approximate planning algorithm.
result Introduces a computationally tractable approximate planning algorithm using state abstraction.
The abstract warns against flawed empirical research in machine learning.
problem Flawed empirical research in machine learning leading to unreliable results.
method Call for more awareness of experimental knowledge plurality and epistemic limitations.
result Current empirical machine learning research should be exploratory, not confirmatory.
Large models collapse epistemic uncertainty, challenging traditional wisdom.
problem Epistemic uncertainty collapse in large models.
method Implicit ensembling and decomposition techniques.
result Larger models can collapse epistemic uncertainty, contrary to expectations.
The study quantifies and compares aleatoric and epistemic discrimination in ML models.
problem Sources of discrimination in ML models and their impact on performance.
method Quantifying aleatoric and epistemic discrimination using statistical experiments and model accuracy.
result State-of-the-art fairness interventions are effective at removing epistemic discrimination but not aleatoric discrimination in datasets with missing values.
This paper simplifies OPE in large state spaces using state abstractions.
problem Accurately evaluating policies offline in large state spaces.
method Developed a backward-model-irrelevance condition and an iterative state abstraction procedure.
result Deeply-abstracted states substantially simplify OPE sample complexity.
Algorithm finds latent structure in value functions for improved reinforcement learning.
problem Finding latent structure in value functions for efficient reinforcement learning.
method Proposes a practical algorithm using two posterior distributions over state abstractions and abstract-state values.
result Substantial performance gains in multi-task settings where tasks share a common, low-dimensional representation.
Tabular Q-Learning with learned state abstractions solves continuous control tasks.
problem Challenging reinforcement learning problems in continuous control.
method Learned state abstraction to transform continuous state-space into discrete.
result Tabular Q-Learning with learned abstractions achieves efficient learning in unseen tasks.
Abstract MDPs enable strategic exploration and fast reward transfer in complex environments.
problem Challenging to learn accurate MDPs for high-dimensional states.
method Learn an abstract MDP over low-dimensional coarse states, using an abstraction function.
result Achieves superhuman performance on Pitfall! and higher reward with fewer samples.
Proposes a new method to measure epistemic uncertainty in Bayesian neural networks.
problem Measuring epistemic uncertainty in Bayesian neural networks for out-of-distribution detection.
method Proposes measuring disagreement between logits and their pre-softmax counterparts as an epistemic uncertainty measure.
result Proposed epistemic uncertainty scores outperform mutual information and equal predictive entropy performance.
Deep neural network learns discrete state abstractions for efficient planning.
problem Efficient sequential decision making in large state spaces.
method Information bottleneck method for learning approximate bisimulations using deep neural encoders and action-conditioned HMM.
result Trained method efficiently plans for unseen goals in multi-goal reinforcement learning.
AIF reformulated as convex MDP for adaptive behavior.
problem Adaptive behavior and policy optimization.
method Formulating AIF as convex MDP, deriving mirror descent algorithm.
result EFE minimization in AIF is equivalent to reward maximization in latent MDP, with epistemic component.
The paper proposes a principle for dynamically adjusting the granularity of reinforcement learning abstractions.
problem Lack of general principles for dynamically adjusting the granularity of reinforcement learning abstractions.
method The paper proposes a principle based on rate-distortion theory, formalized through a performance certificate decomposing value error into learning and abstraction error bounds.
result Soft state-action abstractions can achieve near-optimal performance under substantial lossy compression of state and action information.
DEUA detects diffusion-generated images by accounting for different types of uncertainty.
problem Detecting generated images with varying aleatoric and epistemic uncertainty.
method DEUA framework using Laplace approximation for DEU estimation and asymmetric loss function.
result DEUA achieves state-of-the-art performance on large-scale benchmarks.
VGE provides a practical approach to uncertainty estimation in ensemble models.
problem Uncertainty estimation in ensemble models using additive decomposition breaks down.
method Variance-Gated Ensembles (VGE) introduces a differentiable framework with a signal-to-noise gate.
result VGE provides a Variance-Gated Margin Uncertainty (VGMU) score and Variance-Gated Normalization (VGN) layer.
Proposes method to learn state abstractions that generalize across environments.
problem Learning abstractions that generalize in block MDPs.
method Invariant causal prediction to learn model-irrelevant state abstractions (MISA).
result Proves high probability of outputting a state abstraction corresponding to causal feature set for return.
This paper introduces modal epistemic tools for risk management.
problem Identifying and certifying risk claims when institutions lack the necessary epistemic stance.
method Develops crisp and fuzzy modal semantics for assurance and working commitment, distinguishing between object-level risk claims and meta-level epistemic diagnostics.
result Risk governance should model evidential incompleteness and failures of escalation, not just hazards and losses.
The paper introduces a method to learn Markov state abstractions for reinforcement learning.
problem Learning Markov state representations in complex environments.
method The paper introduces a novel set of conditions and a training procedure combining inverse model estimation and temporal contrastive learning.
result The approach learns representations that capture the underlying structure of the domain and improve sample efficiency.
Second-order methods fail to fully quantify epistemic uncertainty, leading to biased predictions.
problem Incomplete quantification of epistemic uncertainty in machine learning models.
method Analysis of existing second-order uncertainty estimation methods.
result Current methods overestimate aleatoric uncertainty and underestimate epistemic uncertainty, leading to biased predictions.
Bayesian Neural Networks show unexpected collapse of epistemic uncertainty with large models and little data.
problem Unexpected collapse of epistemic uncertainty in Bayesian Neural Networks.
method Experiments with varying model size and training data size.
result Epistemic uncertainty collapses in the presence of large models and sometimes little data.
A new method shapes reinforcement learning environments by abstracting large state spaces.
problem Learning in large, noisy environments with sparse feedback.
method Environment shaping using state abstraction.
result Agent's policy in shaped environment preserves near-optimal behavior in original environment.
Overparametrized neural networks retain significant epistemic uncertainty even with sufficient data.
problem Epistemic uncertainty in overparametrized neural networks persists despite model identifiability.
method Analysis of non-identifiability and characterization of residual uncertainty in one-hidden-layer ReLU networks.
result Substantial parameter uncertainty remains even when the underlying function is fully identified.
DEUP directly predicts epistemic uncertainty, improving model optimization and exploration.
problem Existing measures of epistemic uncertainty do not account for model misspecification.
method Proposes a framework to estimate excess risk as a measure of epistemic uncertainty, using a secondary predictor for generalization error.
result DEUP improves sequential model optimization and exploration in interactive learning environments.
CLEAR calibrates both aleatoric and epistemic uncertainties for better predictive intervals.
problem Balanced uncertainty quantification for reliable predictive modeling.
method CLEAR uses two parameters, γ1 and γ2, to combine aleatoric and epistemic uncertainties.
result Clear achieves significant improvements in interval width and coverage.
New method finds unseen states for RL, improving performance.
problem Offline RL struggles with unseen states and actions.
method Value-informed state perturbations and filtering.
result Improved performance in offline RL tasks.
The standard taxonomy of predictive uncertainty is inconsistent with standard measures.
problem Uncertainty taxonomy and measure inconsistency
method Proof of inconsistency
result Uncertainty is not reducible to data collection
Proposes EVE for efficient exploration in reinforcement learning.
problem Efficient exploration in reinforcement learning.
method EVE: a recipe for posterior over parameters, facilitating efficient exploration.
result Competitive performance on benchmarks, efficient exploration confirmed.
New method estimates model uncertainty in regression.
problem Challenges in distinguishing aleatoric and epistemic uncertainty.
method Conditional predictions with model's initial output.
result Rigorous frequentist approach to epistemic uncertainty.
New method isolates epistemic uncertainty in diffusion models, improving plausibility scores.
problem Uncertainty quantification in diffusion models, especially epistemic uncertainty.
method Fisher information based approach using FLARE (Fisher-Laplace Randomized Estimator).
result FLARE improves uncertainty estimation in synthetic time-series generation tasks.
Introduces epistemic deep learning for better uncertainty estimation in neural networks.
problem Uncertainty quantification in deep neural networks.
method Random-set convolutional neural networks with belief function-based loss functions.
result Epistemic approach produces better performance in uncertainty estimation.
Deep ensembles effectively capture epistemic uncertainty through training stochasticity, providing a frequentist perspective.
problem Understanding and quantifying epistemic uncertainty in machine learning models.
method Bootstrap-based estimator and decomposition of deep ensembles into data variability and training stochasticity.
result Deep ensembles primarily capture training stochasticity, explaining their effectiveness in quantifying epistemic uncertainty.
New framework identifies and reduces errors in machine learning under distribution shift.
problem Errors in machine learning models when distributions change.
method Developed a principled framework to characterize and eliminate epistemic errors in imperfect multitask learning.
result Provided a decompositional epistemic error bound for general settings of distribution shift.
Ensembles of neural networks (NNs) have long been used to estimate predictive uncertainty; a small number of NNs are trained from different initialisations and sometimes on differing versions of the dataset. The variance of the ensemble's predictions is interpreted as its epistemic uncertainty. The appeal of ensembling…
STAR framework reduces OPE variance by distilling complex problems into discrete ARPs.
problem High variance and bias in off-policy evaluation methods.
method STAR framework that includes various OPE estimators and leverages state abstraction.
result Predictions from ARPs estimated from off-policy data are asymptotically correct.
Various strategies for active learning have been proposed in the machine learning literature. In uncertainty sampling, which is among the most popular approaches, the active learner sequentially queries the label of those instances for which its current prediction is maximally uncertain. The predictions as well as the …
Planning methods can solve temporally extended sequential decision making problems by composing simple behaviors. However, planning requires suitable abstractions for the states and transitions, which typically need to be designed by hand. In contrast, model-free reinforcement learning (RL) can acquire behaviors from l…
Generative models use latent abstractions to create images.
problem Understanding how generative models create high-dimensional data like images.
method Developed a theoretical framework using SDE and information theory.
result Diffusion models can be seen as a non-linear filter driven by latent abstractions.
New framework compares credal sets for hypothesis testing with epistemic uncertainty.
problem Comparing distributions with partial ignorance and epistemic uncertainty.
method Credal two-sample testing framework for convex sets of probability measures.
result Direct integration of epistemic uncertainty in hypothesis testing.
The paper introduces new measures for quantifying uncertainty in machine learning.
problem Uncertainty representation and quantification in machine learning.
method Proper scoring rules for aleatoric and epistemic uncertainty quantification.
result Established a natural bridge between credal set and second-order distribution representations of uncertainty.
CreDRO learns credal ensembles via distributionally robust optimization, improving EU quantification.
problem Quantifying predictive epistemic uncertainty in credal models.
method Distributionally robust optimization to capture EU from training randomness and potential distribution shifts.
result Empirically, CreDRO outperforms existing credal methods on various tasks.
CLAPS improves conformal regression by adaptively scaling interval widths based on last-layer Laplace uncertainty.
problem Lack of adaptive interval width scaling in conformal regression for heterogeneous inputs.
method CLAPS uses heteroscedastic last-layer Laplace uncertainty to adaptively scale interval widths, combining aleatoric and epistemic uncertainties.
result CLAPS provides competitive interval efficiency with nominal-level coverage, reducing to aleatoric scaling as epistemic uncertainty decreases.
QUAM improves uncertainty quantification in deep learning models.
problem Estimating epistemic uncertainty in deep learning models.
method QUAM identifies regions with high divergence between predictions and a reference model.
result QUAM has lower approximation error of epistemic uncertainty compared to previous methods.
EPICSCORE improves conformal scores by explicitly accounting for epistemic uncertainty.
problem Overconfident predictions in data-sparse regions due to lack of epistemic uncertainty.
method Model-agnostic approach using Bayesian techniques like Gaussian Processes, Dropout, and Regression Trees.
result Enhanced predictive intervals that adaptively expand in sparse data regions and maintain compact intervals in abundant data.
Loss minimisation fails to capture epistemic uncertainty in second-order predictors.
problem Capturing epistemic uncertainty in machine learning models.
method Analysis of a second-order learner approach using loss minimisation.
result Loss minimisation does not faithfully represent epistemic uncertainty in second-order predictors.
While deep neural networks have become the go-to approach in computer vision, the vast majority of these models fail to properly capture the uncertainty inherent in their predictions. Estimating this predictive uncertainty can be crucial, for example in automotive applications. In Bayesian deep learning, predictive unc…
The volume of a credal set correlates with epistemic uncertainty in binary classification but not in multi-class.
problem Representing and quantifying epistemic uncertainty in machine learning.
method Examined the geometric representation of credal sets as d-dimensional polytopes and their volume as a measure of uncertainty. result The volume of a credal set is a meaningful measure of epistemic uncertainty in binary classification but not in multi-class.
We present an algorithm, HOMER, for exploration and reinforcement learning in rich observation environments that are summarizable by an unknown latent state space. The algorithm interleaves representation learning to identify a new notion of kinematic state abstraction with strategic exploration to reach new states usi…
GPNs use unlabeled data to estimate uncertainty in Bayesian problems.
problem Limited training data in high-dimensional problems.
method Generative Posterior Networks (GPNs) that approximate the posterior distribution using unlabeled data.
result GPNs improve epistemic uncertainty estimation and scalability.
Unified framework for hierarchical image classification with epistemic uncertainty.
problem Overconfident predictions and lack of logical consistency in deep learning models.
method Neurosymbolic approach with epistemic deep learning, using focal set reasoning and differentiable fuzzy logic.
result Maintains accuracy on par with transformer baselines while providing more calibrated and interpretable predictions.