Singapore

While existing formal verification tools are increasingly effective at verifying local robustness properties, more complex ones, such as those involving multiple neural networks, remain beyond their capabilities: these properties cannot currently be expressed in their specification languages, nor can they be directly verified. We present CAISAR, an open-source platform for specifying and verifying properties of machine learning models, with particular focus on neural networks and support vector machines. CAISAR provides a high-level language for specifying complex properties and integrates several state-of-the-art verifiers for their automatic verification. Through concrete use cases, we show how CAISAR leverages automated graph-editing techniques to translates high-level specifications into queries for the supported verifiers, bridging the embedding gap between user specifications and the corresponding ones that are actually verified.

AAAI 2026

The CAISAR Platform: Extending the Reach of Machine Learning Specification and Verification

formal specification and verification

machine learning

workshop paper

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

The purpose of the AAAI conference series is to promote research in Artificial Intelligence (AI) and foster scientific exchange between researchers, practitioners, scientists, students, and engineers across the entirety of AI and its affiliated disciplines. AAAI-26 will feature technical paper presentations, special tracks, invited speakers, workshops, tutorials, poster sessions, senior member presentations, competitions, and exhibit programs, and a range of other activities to be announced.<br><br>

To access this event page, you need to log in with the **email address you registered with**. <br>Access credentials will be sent to your email from Underline -  subject line "Welcome to AAAI 2026". Please be sure to check your spam email folder if you do not see an email confirmation right away.

Please log in

To access this event page, you are required to register.
Please complete your registration to continue.

We recommend reading [**the registration information**](https://aaai.org/conference/aaai/aaai-26/registration/) first.

**Online Registration Form**: https://aaai.getregistered.net/conference-2026 

Registration Required

We are pleased to announce the Fortieth AAAI Conference on Artificial Intelligence (AAAI-26), which will be held in Singapore EXPO from January 20 to January 27, 2026.

SHapley Additive exPlanations (SHAP) is a key tool for interpreting decision tree ensembles by assigning contribution values to features. The most accurate variant, Background SHAP, uses a background dataset to estimate feature distributions. We introduce WOODELF, a SHAP algorithm that unifies decision trees, game theory, and Boolean logic within a single framework. For each instance, WOODELF constructs a pseudo-Boolean formula that captures their feature values, the structure of the decision tree ensemble, and the entire background dataset. It then leverages this representation to compute Background SHAP in linear time. WOODELF can also compute Shapley interaction values, Banzhaf values, and Banzhaf interaction values. WOODELF runs efficiently on both CPU and GPU. We demonstrate these advantages empirically on two large, commonly used datasets, where WOODELF achieves at least an order-of-magnitude speedup over state-of-the-art methods.

From Decision Trees to Boolean Logic: A Fast and Unified SHAP Algorithm

Case-based reasoning networks are models that make predictions based on similarity between the input and prototypical parts of training samples. Such models explain each decision by pointing to the prototypes that contributed the most to the final outcome. While promising, such explanations are sometimes misleading, which hampers their usefulness in safety-critical contexts. Several instances may lead to different predictions and yet have the same explanation. Drawing inspiration from formal eXplainable AI, we propose Abductive Latent Explanations (ALEs), a formalism to express sufficient conditions on the latent representation of the instance that imply the prediction. Our approach combines the inherent interpretability of case-based reasoning models and the guarantees provided by formal XAI. We propose a solver-free algorithm for generating ALEs based on three distinct paradigms, compare them, and present the feasibility of our approach on diverse datasets for image classification.

Formal Abductive Latent Explanations for Prototype-Based Networks

In reinforcement learning (RL), it is advantageous to consider additional constraints on the action space to ensure safety. Existing work on such action-constrained RL often lacks expressive policy updates, computational efficiency, and predictable runtime. Recent work proposes to use truncated normal distributions for stochastic policy gradient methods. However, the computation of key characteristics, such as the entropy, log-probability, and their gradients, becomes intractable. Hence, prior work approximates these using the non-truncated distributions, which severely degrades performance. We argue that accurate estimation of these characteristics is crucial in the action-constrained RL setting, and propose efficient numerical approximations for them. We also provide an efficient sampling strategy for truncated policy distributions and validate our approach on three benchmark environments, which demonstrate significant performance improvements when using accurate estimations.

Improving Stochastic Action-Constrained Reinforcement Learning via Truncated Distributions

Robust Markov Decision Processes generalize classical MDPs that consider uncertainties in transition probabilities by defining a set of possible transition functions. An objective is a set of runs (or infinite trajectories) of the RMDP, and the value for an objective is the maximal probability that the agent can guarantee against the adversarial environment. We consider (a) reachability objectives, where given a target set of states, the goal is to eventually arrive at one of them; and (b) parity objectives, which are a canonical representation for $\omega$-regular objectives. In this work, we study the qualitative problem for reachability and parity objectives on RMDPs without making any assumption over the structures of the RMDPs, e.g., unichain or aperiodic. We present efficient algorithms with oracle access to uncertainty sets that solve qualitative problems of reachability and parity objectives, together with experimental results.

Qualitative Analysis of $\omega$-Regular Objectives on Robust MDPs

Python has become the dominant language for general-purpose programming, yet it lacks robust tools for formal verification. In contrast, programmers working in languages such as C benefit from mature model checkers, for example CBMC, which enable exhaustive symbolic reasoning and fault localisation. In this paper, we propose PyVeritas, a novel framework that leverages Large Language Models (LLMs) for high-level transpilation from Python to C, followed by bounded model checking and MaxSAT-based fault localisation in the generated C code. PyVeritas enables verification and bug localisation for Python code using existing model checking tools for C. Our empirical evaluation demonstrates that LLM-based transpilation can achieve a high degree of accuracy, up to 90% for some LLMs, enabling effective development environment that supports assertion-based verification and interpretable fault diagnosis for small yet non-trivial Python programs.

PyVeritas: On Verifying Python via LLM-Based Transpilation and Bounded Model Checking for C

Privacy leakage in AI-based decision processes poses significant risks, particularly when sensitive information can be inferred. We propose a formal framework to audit privacy leakage using abductive explanations, which identifies minimal sufficient evidence justifying model decisions and determines whether sensitive information disclosed. Our framework formalizes both individual and system-level leakage, introducing the notion of Potentially Applicable Explanations (PAE) to identify individuals whose outcomes can shield those with sensitive features. This approach provides rigorous privacy guarantees while producing human-understandable explanations, a key requirement for auditing tools. Experimental evaluation on the German Credit Dataset illustrates how the importance of sensitive literal in the model decision process affects privacy leakage.

Beyond Verification: Abductive Explanations for Post-AI Assessment of Privacy Leakage

We explore AI-driven distributed-systems policy design by combining stochastic code generation from large language models (LLMs) with deterministic verification in a domain-specific simulator. Using a Function-as-a-Service runtime (Bauplan) and its open-source simulator (Eudoxia) as a case study, we frame scheduler design as an iterative generate-and-verify loop: an LLM proposes a Python policy, the simulator evaluates it on standardized traces, and structured feedback steers subsequent generations. This setup preserves interpretability while enabling targeted search over a large design space. We detail the system architecture and report preliminary results on throughput improvements across multiple models. Beyond early gains, we discuss the limits of the current setup and outline next steps; in particular, we conjecture that AI will be crucial for scaling this methodology by helping to bootstrap new simulators.

AI for Distributed Systems Design: Scalable Cloud Optimization Through Repeated LLMs Sampling And Simulators

We introduce INDIMATHBENCH, a human-verified benchmark designed to evaluate mathematical theorem proving, curated using an AI-powered human-assisted pipeline for formalizing natural language problems in Lean. INDIMATHBENCH is composed of 312 formal Lean 4 theorems paired with their corresponding informal problem statements, sourced from Indian Mathematics Olympiads. Through category-based retrieval, iterative compiler feedback, and multi-model ensembles, our pipeline generates candidate formalizations that experts efficiently validate via an interactive dashboard with automated quality summaries. Evaluation across multiple frontier models demonstrates that autoformalization remains challenging, with substantial gaps between syntactic validity and semantic correctness, while theorem proving success rates remain low even with iterative refinement, demonstrating that INDIMATHBENCH presents a challenging testbed for mathematical reasoning.

IndiMathBench: Autoformalizing Mathematical Reasoning Problems with a Human Touch

Reinforcement learning (RL) commonly relies on scalar rewards with limited ability to express temporal, conditional, or safety-critical goals, and can lead to reward hacking. Temporal logic expressible via the more general class of $\omega$-regular objectives addresses this by precisely specifying rich behavioural properties. We address both limitations simulatenously by combining $\omega$-regular objectives with explicit constraints, allowing safety requirements and optimisation targets to be treated separately. We develop a model-based RL algorithm based on linear programming, which in the limit produces a policy maximizing the probability of satisfying an $\omega$-regular objective while also adhering to $\omega$-regular constraints within specified thresholds. Furthermore, we establish a translation to constrained limit-average problems with optimality-preserving guarantees.

Reinforcement Learning with $\omega$-Regular Objectives and Constraints

Clinical and regulatory discussions about trustworthy therapeutic AI speak in ethical and legal terms, while technical work reports performance through AUC, F1, PPV, or survival concordance. The mapping between these numerical summaries and the social norms they support is implicit and often incoherent. This gap is acute for therapeutic AI systems built on spatially resolved data---such as digital pathology, radiology, and spatial omics---where models guide target selection, biomarker discovery, and responder prediction. We provide a technical--normative analysis of common evaluation metrics. We formalize metric families by their aggregation operations: expectation-based (expected loss; AUC as a U-statistic over pairs), quantile/tail-based (median, upper quantiles, CVaR), supremum-type (worst-group risk, minimax regret), thresholded confusion-matrix ratios (PPV/precision, sensitivity, F1), and ranking metrics (top- 
, average precision). For each family we identify the implicit social norm: maximizing average benefit, protecting typical patients, guarding against worst-case harms, or prioritizing top-predicted benefit. We prove an incompatibility result showing that high AUC can coexist with very low worst-group sensitivity under deployment-relevant thresholding, and we illustrate further incompatibilities via clinical examples. We then propose a metric design framework: (i) explicit normative declarations, (ii) multi-objective evaluation with subgroup and tail-risk constraints, and (iii) deployment checklists tying thresholds to institutional responsibilities. The goal is not to replace ethical debate with formulas, but to make explicit and auditable the value judgments encoded by metric choices.

Content not yet available

Next from AAAI 2026

From Decision Trees to Boolean Logic: A Fast and Unified SHAP Algorithm

Stay up to date with the latest Underline news!

PRESENTATIONS

CONFERENCES

COMPANY

RESOURCES