Multimodal large reasoning models (MLRMs) are rapidly advancing vision-language reasoning and are emerging as a foundation for cross-modal intelligence. Hallucination remains a persistent failure mode, manifesting itself as erroneous reasoning chains and misinterpretation of visual content. In this study, we observe that attention heads exhibit a staged division: shallow heads predominantly serve perception, while deeper heads shift toward symbolic reasoning, revealing two major causes of hallucination, namely perceptual bias and reasoning drift. To address these issues, we propose a lightweight and interpretable two-step plugin, Functional Head Identification and Class-conditioned Rescaling, which locates perception- and reasoning-oriented heads and regulates their contributions without retraining. Evaluations on three real-world MLRMs (Kimi-VL, Ocean-R1, R1-Onevision), six benchmarks across three domains, and four baselines show that our plugin achieves an average improvement of 5% and up to 15%, with only <1% additional computation and 9% of baseline latency. Our approach is completely model-agnostic and significantly enhances both the reliability and interpretability of the off-the-shelf MLRMs, thereby enabling their safe deployment in high-stakes applications. Our code is available at https://anonymous.4open.science/r/Functional-Attention-Control.

This work presents a formal framework for quantifying the internal dependencies between functional subsystems within artificial agents whose belief states are composed of structured linguistic fragments. Building on the Semantic Manifold framework, which organizes belief content into functional sectors and stratifies them across hierarchical levels of abstraction, we introduce a system of sectoral coupling constants that characterize how one cognitive sector influences another within a fixed level of abstraction. The complete set of these constants forms an agent-specific coupling profile that governs internal information flow, shaping the agent's overall processing tendencies and cognitive style. We provide a detailed taxonomy of these intra-level coupling roles, covering domains such as perceptual integration, memory access and formation, planning, meta-cognition, execution control, and affective modulation. We also explore how these coupling profiles generate feedback loops, systemic dynamics, and emergent signatures of cognitive behavior. Methodologies for inferring these profiles from behavioral or internal agent data are outlined, along with a discussion of how these couplings evolve across abstraction levels. This framework contributes a mechanistic and interpretable approach to modeling complex cognition, with applications in AI system design, alignment diagnostics, and the analysis of emergent agent behavior.

Code generation with large language models has shown significant promise, especially when employing retrieval-augmented generation (RAG) with few-shot examples. However, selecting effective examples that enhance generation quality remains a challenging task, particularly when the target programming language (PL) is underrepresented. In this study, we present two key findings: (1) retrieving examples whose presented algorithmic plans can be referenced for generating the desired behavior significantly improves generation accuracy, and (2) converting code into pseudocode effectively captures such algorithmic plans, enhancing retrieval quality even when the source and the target PLs are different. Based on these findings, we propose Plan-as-query Example Retrieval for few-shot prompting in Code generation (PERC), a novel framework that utilizes algorithmic plans to identify and retrieve effective examples. We validate the effectiveness of PERC through extensive experiments on the CodeContests, HumanEval and MultiPL-E benchmarks: PERC consistently outperforms the state-of-the-art RAG methods in code generation, both when the source and target programming languages match or differ, highlighting its adaptability and robustness in diverse coding environments.

We propose selective debiasing -- an inference-time safety mechanism that aims to increase the overall quality of models in terms of prediction performance and fairness in the situation when re-training a model is prohibitive. The method is inspired by selective prediction, where some predictions that are considered low quality are discarded at inference time. In our approach, we identify the potentially biased model predictions and, instead of discarding them, we debias them using LEACE -- a post-processing debiasing method. To select problematic predictions, we propose a bias quantification approach based on KL divergence, which achieves better results than standard UQ methods. Experiments with text classification datasets demonstrate that selective debiasing helps to close the performance gap between post-processing methods and at-training and pre-processing debiasing techniques.

Large vision-language models (LVLMs) have recently dramatically pushed the state of the art in image captioning and many image understanding tasks (e.g., visual question answering). LVLMs, however, often \textit{hallucinate} and produce captions that mention concepts that cannot be found in the image. These hallucinations erode the trustworthiness of LVLMs and are arguably among the main obstacles to their ubiquitous adoption. Recent work suggests that addition of grounding objectives -- those that explicitly align image regions or objects to text spans -- reduces the amount of LVLM hallucination. Although intuitive, this claim is not empirically justified as the reduction effects have been established, we argue, with flawed evaluation protocols that (i) rely on data (i.e., MSCOCO) that has been extensively used in LVLM training and (ii) measure hallucination via question answering rather than open-ended caption generation. In this work, in contrast, we offer the first systematic analysis of the effect of fine-grained object grounding on LVLM hallucination under an evaluation protocol that more realistically captures LVLM hallucination in open generation. Our extensive experiments over three backbone LLMs reveal that grounding objectives have little to no effect on object hallucination in open caption generation.

Fairness in AI is a growing concern for high-stakes decision making. Engaging stakeholders, especially lay users, in fair AI development is promising yet overlooked. Recent efforts explore enabling lay users to provide AI fairness-related feedback, but there is still a lack of understanding of how to integrate users' feedback into an AI model and the impacts of doing so. To bridge this gap, we collected feedback from 58 lay users on the fairness of a XGBoost model trained on the Home Credit dataset, and conducted offline experiments to investigate the effects of retraining models on accuracy, and individual and group fairness. Our work contributes baseline results of integrating user fairness feedback in XGBoost, and a dataset and code framework to bootstrap research in engaging stakeholders in AI fairness. Our discussion highlights the challenges of employing user feedback in AI fairness and points the way to a future application area of interactive machine learning.

Leveraging machine-learning methods to predict outcomes on some unlabeled datasets and then using these pseudo-outcomes in subsequent statistical inference is common in modern data analysis. Inference in this setting is often called post-prediction inference. We propose a novel, assumption-lean framework for inference under post-prediction setting, called \emph{Prediction De-Correlated inference} (PDC). Our approach can automatically adapt to any black-box machine-learning model and consistently outperforms supervised methods. The PDC framework also offers easy extensibility for accommodating multiple predictive models. Both numerical results and real-world data analysis support our theoretical results.

The necessity of data driven decisions in healthcare strategy formulation is rapidly increasing. A reliable framework which helps identify factors impacting a Healthcare Provider Facility or a Hospital (from here on termed as Facility) Market Share is of key importance. This pilot study aims at developing a data driven Machine Learning - Regression framework which aids strategists in formulating key decisions to improve the Facilitys Market Share which in turn impacts in improving the quality of healthcare services. The US (United States) healthcare business is chosen for the study; and the data spanning across 60 key Facilities in Washington State and about 3 years of historical data is considered. In the current analysis Market Share is termed as the ratio of facility encounters to the total encounters among the group of potential competitor facilities. The current study proposes a novel two-pronged approach of competitor identification and regression approach to evaluate and predict market share, respectively. Leveraged model agnostic technique, SHAP, to quantify the relative importance of features impacting the market share. The proposed method to identify pool of competitors in current analysis, develops Directed Acyclic Graphs (DAGs), feature level word vectors and evaluates the key connected components at facility level. This technique is robust since its data driven which minimizes the bias from empirical techniques. Post identifying the set of competitors among facilities, developed Regression model to predict the Market share. For relative quantification of features at a facility level, incorporated SHAP a model agnostic explainer. This helped to identify and rank the attributes at each facility which impacts the market share.

Today's robots often interface with data-driven perception and planning models with classical model-predictive controllers (MPC). Often, such learned perception/planning models produce erroneous waypoint predictions on out-of-distribution (OoD) or even adversarial visual inputs, which increase control costs. However, today's methods to train robust perception models are largely task-agnostic - they augment a dataset using random image transformations or adversarial examples targeted at the vision model in isolation. As such, they often introduce pixel perturbations that are ultimately benign for control. In contrast to prior work that synthesizes adversarial examples for single-step vision tasks, our key contribution is to synthesize adversarial scenarios tailored to multi-step, model-based control. To do so, we use differentiable MPC methods to calculate the sensitivity of a model-based controller to errors in state estimation. We show that re-training vision models on these adversarial datasets improves control performance on OoD test scenarios by up to 36.2% compared to standard task-agnostic data augmentation. We demonstrate our method on examples of robotic navigation, manipulation in RoboSuite, and control of an autonomous air vehicle.