Probabilistic classifiers are central for making informed decisions under uncertainty. Based on the maximum expected utility principle, optimal decision rules can be derived using the posterior class probabilities and misclassification costs. Yet, in practice only learned approximations of the oracle posterior probabilities are available. In this work, we quantify the excess risk (a.k.a. regret) incurred using approximate posterior probabilities in batch binary decision-making. We provide analytical expressions for miscalibration-induced regret ($R^{CL}$), as well as tight and informative upper and lower bounds on the regret of calibrated classifiers ($R^{GL}$). These expressions allow us to identify regimes where recalibration alone addresses most of the regret, and regimes where the regret is dominated by the grouping loss, which calls for post-training beyond recalibration. Crucially, both $R^{CL}$ and $R^{Gl}$ can be estimated in practice using a calibration curve and a recent grouping loss estimator. On NLP experiments, we show that these quantities identify when the expected gain of more advanced post-training is worth the operational cost. Finally, we highlight the potential of multicalibration approaches as efficient alternatives to costlier fine-tuning approaches.

Missing values are prevalent across various fields, posing challenges for training and deploying predictive models. In this context, imputation is a common practice, driven by the hope that accurate imputations will enhance predictions. However, recent theoretical and empirical studies indicate that simple constant imputation can be consistent and competitive. This empirical study aims at clarifying *if* and *when* investing in advanced imputation methods yields significantly better predictions. Relating imputation and predictive accuracies across combinations of imputation and predictive models on 19 datasets, we show that imputation accuracy matters less i) when using expressive models, ii) when incorporating missingness indicators as complementary inputs, iii) matters much more for generated linear outcomes than for real-data outcomes. Interestingly, we also show that the use of the missingness indicator is beneficial to the prediction performance, even in MCAR scenarios. Overall, on real-data with powerful models, imputation quality has only a minor effect on prediction performance. Thus, investing in better imputations for improved predictions often offers limited benefits.

The long-standing dominance of gradient-boosted decision trees on tabular data is currently challenged by tabular foundation models using In-Context Learning (ICL): setting the training data as context for the test data and predicting in a single forward pass without parameter updates. While TabPFNv2 foundation model excels on tables with up to 10K samples, its alternating column- and row-wise attentions make handling large training sets computationally prohibitive. So, can ICL be effectively scaled and deliver a benefit for larger tables? We introduce TabICL, a tabular foundation model for classification, pretrained on synthetic datasets with up to 60K samples and capable of handling 500K samples on affordable resources. This is enabled by a novel two-stage architecture: a column-then-row attention mechanism to build fixed-dimensional embeddings of rows, followed by a transformer for efficient ICL. Across 200 classification datasets from the TALENT benchmark, TabICL is on par with TabPFNv2 while being systematically faster (up to 10 times), and significantly outperforms all other approaches. On 53 datasets with over 10K samples, TabICL surpasses both TabPFNv2 and CatBoost, demonstrating the potential of ICL for large data. Pretraining code, inference code, and pre-trained models are available at https://github.com/soda-inria/tabicl.

How to learn a good predictor on data with missing values? Most efforts focus on first imputing as well as possible and second learning on the completed data to predict the outcome. Yet, this widespread practice has no theoretical grounding. Here we show that for almost all imputation functions, an impute-then-regress procedure with a powerful learner is Bayes optimal. This result holds for all missing-values mechanisms, in contrast with the classic statistical results that require missing-at-random settings to use imputation in probabilistic modeling. Moreover, it implies that perfect conditional imputation is not needed for good prediction asymptotically. In fact, we show that on perfectly imputed data the best regression function will generally be discontinuous, which makes it hard to learn. Crafting instead the imputation so as to leave the regression function unchanged simply shifts the problem to learning discontinuous imputations. Rather, we suggest that it is easier to learn imputation and regression jointly. We propose such a procedure, adapting NeuMiss, a neural network capturing the conditional links across observed and unobserved variables whatever the missing-value pattern. Our experiments confirm that joint imputation and regression through NeuMiss is better than various two step procedures in a finite-sample regime.

We consider building predictors when the data have missing values. We study the seemingly-simple case where the target to predict is a linear function of the fully observed data and we show that, in the presence of missing values, the optimal predictor is not linear in general. In the particular Gaussian case, it can be written as a linear function of multiway interactions between the observed data and the various missing value indicators. Due to its intrinsic complexity, we study a simple approximation and prove generalization bounds with finite samples, highlighting regimes for which each method performs best. We then show that multilayer perceptrons with ReLU activation functions can be consistent, and can explore good trade-offs between the true model and approximations. Our study highlights the interesting family of models that are beneficial to fit with missing values depending on the amount of data available.

The presence of missing values makes supervised learning much more challenging. Indeed, previous work has shown that even when the response is a linear function of the complete data, the optimal predictor is a complex function of the observed entries and the missingness indicator. As a result, the computational or sample complexities of consistent approaches depend on the number of missing patterns, which can be exponential in the number of dimensions. In this work, we derive the analytical form of the optimal predictor under a linearity assumption and various missing data mechanisms including Missing at Random (MAR) and self-masking (Missing Not At Random). Based on a Neumann-series approximation of the optimal predictor, we propose a new principled architecture, named NeuMiss networks. Their originality and strength come from the use of a new type of non-linearity: the multiplication by the missingness indicator. We provide an upper bound on the Bayes risk of NeuMiss networks, and show that they have good predictive accuracy with both a number of parameters and a computational complexity independent of the number of missing data patterns. As a result they scale well to problems with many features, and remain statistically efficient for medium-sized samples. Moreover, we show that, contrary to procedures using EM or imputation, they are robust to the missing data mechanism, including difficult MNAR settings such as self-masking.

Learning sparse linear models with two-way interactions is desirable in many application domains such as genomics. $\ell_1$-regularised linear models are popular to estimate sparse models, yet standard implementations fail to address specifically the quadratic explosion of candidate two-way interactions in high dimensions, and typically do not scale to genetic data with hundreds of thousands of features. Here we present WHInter, a working set algorithm to solve large $\ell_1$-regularised problems with two-way interactions for binary design matrices. The novelty of WHInter stems from a new bound to efficiently identify working sets while avoiding to scan all features, and on fast computations inspired from solutions to the maximum inner product search problem. We apply WHInter to simulated and real genetic data and show that it is more scalable and two orders of magnitude faster than the state of the art.

Genome-wide somatic mutation profiles of tumours can now be assessed efficiently and promise to move precision medicine forward. Statistical analysis of mutation profiles is however challenging due to the low frequency of most mutations, the varying mutation rates across tumours, and the presence of a majority of passenger events that hide the contribution of driver events. Here we propose a method, NetNorM, to represent whole-exome somatic mutation data in a form that enhances cancer-relevant information using a gene network as background knowledge. We evaluate its relevance for two tasks: survival prediction and unsupervised patient stratification. Using data from 8 cancer types from The Cancer Genome Atlas (TCGA), we show that it improves over the raw binary mutation data and network diffusion for these two tasks. In doing so, we also provide a thorough assessment of somatic mutations prognostic power which has been overlooked by previous studies because of the sparse and binary nature of mutations.