We present an information-theoretic framework for discrete diffusion models that yields principled estimators of log-likelihood using score-matching losses. Inspired by the I-MMSE identity for the Gaussian setup, we derive analogous results for the discrete setting. Specifically, we introduce the Information-Minimum Denoising Score Entropy (I-MDSE) relation, which links mutual information between data and its diffused version to the minimum denoising score entropy (DSE) loss. We extend this theory to masked diffusion and establish the Information-Minimum Denoising Cross-Entropy (I-MDCE) relation, connecting cross-entropy losses to mutual information in discrete masked processes. These results provide a time-integral decomposition of the log-likelihood of the data in terms of optimal score-based losses, showing that commonly used losses such as DSE and DCE are not merely variational bounds but tight and principled estimators of log-likelihood. The I-MDCE decomposition further enables practical extensions, including time-free formula, conditional likelihood estimation in prompt-response tasks, and coupled Monte Carlo estimation of likelihood ratios. Experiments on synthetic and real-world data confirm the accuracy, variance stability, and utility of our estimators. The code is publicly available at https://github.com/Dongjae0324/infodis.

Probabilistic predictions are at the heart of modern data science.

Calibration is a classical notion from the forecasting literature which aims to address the question: how should predicted probabilities be interpreted? In a world where we only get to observe (discrete) outcomes, how should we evaluate a predictor that hypothesizes (continuous) probabilities over possible outcomes? The study of calibration has seen a surge of recent interest, given the ubiquity of probabilistic predictions in machine learning. This survey describes recent work on the foundational questions of how to define and measure calibration error, and what these measures mean for downstream decision makers who wish to use the predictions to make decisions. A unifying viewpoint that emerges is that of calibration as a form of indistinguishability, between the world hypothesized by the predictor and the real world (governed by nature or the Bayes optimal predictor). In this view, various calibration measures quantify the extent to which the two worlds can be told apart by certain classes of distinguishers or statistical measures.

Model updates (new hyperparameters, kernels, depths, solvers, or data) change performance, but the \emph{reason} often remains opaque. We introduce \textbf{Delta-Attribution} (\mbox{$Δ$-Attribution}), a model-agnostic framework that explains \emph{what changed} between versions $A$ and $B$ by differencing per-feature attributions: $Δϕ(x)=ϕ_B(x)-ϕ_A(x)$. We evaluate $Δϕ$ with a \emph{$Δ$-Attribution Quality Suite} covering magnitude/sparsity (L1, Top-$k$, entropy), agreement/shift (rank-overlap@10, Jensen--Shannon divergence), behavioural alignment (Delta Conservation Error, DCE; Behaviour--Attribution Coupling, BAC; CO$Δ$F), and robustness (noise, baseline sensitivity, grouped occlusion). Instantiated via fast occlusion/clamping in standardized space with a class-anchored margin and baseline averaging, we audit 45 settings: five classical families (Logistic Regression, SVC, Random Forests, Gradient Boosting, $k$NN), three datasets (Breast Cancer, Wine, Digits), and three A/B pairs per family. \textbf{Findings.} Inductive-bias changes yield large, behaviour-aligned deltas (e.g., SVC poly$\!\rightarrow$rbf on Breast Cancer: BAC$\approx$0.998, DCE$\approx$6.6; Random Forest feature-rule swap on Digits: BAC$\approx$0.997, DCE$\approx$7.5), while ``cosmetic'' tweaks (SVC \texttt{gamma=scale} vs.\ \texttt{auto}, $k$NN search) show rank-overlap@10$=1.0$ and DCE$\approx$0. The largest redistribution appears for deeper GB on Breast Cancer (JSD$\approx$0.357). $Δ$-Attribution offers a lightweight update audit that complements accuracy by distinguishing benign changes from behaviourally meaningful or risky reliance shifts.

Forecast probabilities often serve as critical inputs for binary decision making. In such settings, calibration$\unicode{x2014}$ensuring forecasted probabilities match empirical frequencies$\unicode{x2014}$is essential. Although the common notion of Expected Calibration Error (ECE) provides actionable insights for decision making, it is not testable: it cannot be empirically estimated in many practical cases. Conversely, the recently proposed Distance from Calibration (dCE) is testable but is not actionable since it lacks decision-theoretic guarantees needed for high-stakes applications. We introduce Cutoff Calibration Error, a calibration measure that bridges this gap by assessing calibration over intervals of forecasted probabilities. We show that Cutoff Calibration Error is both testable and actionable and examine its implications for popular post-hoc calibration methods, such as isotonic regression and Platt scaling.

Rigorous evaluation of the causal effects of semantic features on language model predictions can be hard to achieve for natural language reasoning problems. However, this is such a desirable form of analysis from both an interpretability and model evaluation perspective, that it is valuable to investigate specific patterns of reasoning with enough structure and regularity to identify and quantify systematic reasoning failures in widely-used models. In this vein, we pick a portion of the NLI task for which an explicit causal diagram can be systematically constructed: the case where across two sentences (the premise and hypothesis), two related words/terms occur in a shared context. In this work, we apply causal effect estimation strategies to measure the effect of context interventions (whose effect on the entailment label is mediated by the semantic monotonicity characteristic) and interventions on the inserted word-pair (whose effect on the entailment label is mediated by the relation between these words). Extending related work on causal analysis of NLP models in different settings, we perform an extensive interventional study on the NLI task to investigate robustness to irrelevant changes and sensitivity to impactful changes of Transformers. The results strongly bolster the fact that similar benchmark accuracy scores may be observed for models that exhibit very different behaviour. Moreover, our methodology reinforces previously suspected biases from a causal perspective, including biases in favour of upward-monotone contexts and ignoring the effects of negation markers.

In the recent literature on machine learning and decision making, calibration has emerged as a desirable and widely-studied statistical property of the outputs of binary prediction models. However, the algorithmic aspects of measuring model calibration have remained relatively less well-explored. Motivated by [BGHN23], which proposed a rigorous framework for measuring distances to calibration, we initiate the algorithmic study of calibration through the lens of property testing. We define the problem of calibration testing from samples where given $n$ draws from a distribution $\mathcal{D}$ on (predictions, binary outcomes), our goal is to distinguish between the case where $\mathcal{D}$ is perfectly calibrated, and the case where $\mathcal{D}$ is $\varepsilon$-far from calibration. We design an algorithm based on approximate linear programming, which solves calibration testing information-theoretically optimally (up to constant factors) in time $O(n^{1.5} \log(n))$. This improves upon state-of-the-art black-box linear program solvers requiring $\Omega(n^\omega)$ time, where $\omega > 2$ is the exponent of matrix multiplication. We also develop algorithms for tolerant variants of our testing problem, and give sample complexity lower bounds for alternative calibration distances to the one considered in this work. Finally, we present preliminary experiments showing that the testing problem we define faithfully captures standard notions of calibration, and that our algorithms scale to accommodate moderate sample sizes.

Evaluating the quality and variability of text generated by Large Language Models (LLMs) poses a significant, yet unresolved research challenge. Traditional evaluation methods, such as ROUGE and BERTScore, which measure token similarity, often fail to capture the holistic semantic equivalence. This results in a low correlation with human judgments and intuition, which is especially problematic in high-stakes applications like healthcare and finance where reliability, safety, and robust decision-making are highly critical. This work proposes DCR, an automated framework for evaluating and improving the consistency of LLM-generated texts using a divide-conquer-reasoning approach. Unlike existing LLM-based evaluators that operate at the paragraph level, our method employs a divide-and-conquer evaluator (DCE) that breaks down the paragraph-to-paragraph comparison between two generated responses into individual sentence-to-paragraph comparisons, each evaluated based on predefined criteria. To facilitate this approach, we introduce an automatic metric converter (AMC) that translates the output from DCE into an interpretable numeric score. Beyond the consistency evaluation, we further present a reason-assisted improver (RAI) that leverages the analytical reasons with explanations identified by DCE to generate new responses aimed at reducing these inconsistencies. Through comprehensive and systematic empirical analysis, we show that our approach outperforms state-of-the-art methods by a large margin (e.g., +19.3% and +24.3% on the SummEval dataset) in evaluating the consistency of LLM generation across multiple benchmarks in semantic, factual, and summarization consistency tasks. Our approach also substantially reduces nearly 90% of output inconsistencies, showing promise for effective hallucination mitigation.

Calibration measures and reliability diagrams are two fundamental tools for measuring and interpreting the calibration of probabilistic predictors. Calibration measures quantify the degree of miscalibration, and reliability diagrams visualize the structure of this miscalibration. However, the most common constructions of reliability diagrams and calibration measures -- binning and ECE -- both suffer from well-known flaws (e.g. discontinuity). We show that a simple modification fixes both constructions: first smooth the observations using an RBF kernel, then compute the Expected Calibration Error (ECE) of this smoothed function. We prove that with a careful choice of bandwidth, this method yields a calibration measure that is well-behaved in the sense of (B{\l}asiok, Gopalan, Hu, and Nakkiran 2023a) -- a consistent calibration measure. We call this measure the SmoothECE. Moreover, the reliability diagram obtained from this smoothed function visually encodes the SmoothECE, just as binned reliability diagrams encode the BinnedECE. We also provide a Python package with simple, hyperparameter-free methods for measuring and plotting calibration: `pip install relplot\`.