arXiv subjects

Artificial intelligence

Search indexed arXiv papers on artificial intelligence and machine learning, including cs.AI metadata. Follow the original manuscripts for methods, experiments and version history.

Follow this topic with RSS

At least 793 records · Page 44Linked to original sources

arXiv2011

An Algebraic Graphical Model for Decision with Uncertainties, Feasibilities, and Utilities

Numerous formalisms and dedicated algorithms have been designed in the last decades to model and solve decision making problems. Some formalisms, such as constraint networks, can express "simple" decision problems, while others are designed to take into account uncertainties, unfeasible decisions, and utilities. Even in a single formalism, several variants are often proposed to model different types of uncertainty (probability, possibility...) or utility (additive or not). In this article, we introduce an algebraic graphical model that encompasses a large number of such formalisms: (1) we first adapt previous structures from Friedman, Chu and Halpern for representing uncertainty, utility, and expected utility in order to deal with generic forms of sequential decision making; (2) on these structures, we then introduce composite graphical models that express information via variables linked by "local" functions, thanks to conditional independence; (3) on these graphical models, we finally define a simple class of queries which can represent various scenarios in terms of observabilities and controllabilities. A natural decision-tree semantics for such queries is completed by an equivalent operational semantics, which induces generic algorithms. The proposed framework, called the Plausibility-Feasibility-Utility (PFU) framework, not only provides a better understanding of the links between existing formalisms, but it also covers yet unpublished frameworks (such as possibilistic influence diagrams) and unifies formalisms such as quantified boolean formulas and influence diagrams. Our backtrack and variable elimination generic algorithms are a first step towards unified algorithms.

cs.AI↗

arXiv2011

Semantic Matchmaking as Non-Monotonic Reasoning: A Description Logic Approach

Matchmaking arises when supply and demand meet in an electronic marketplace, or when agents search for a web service to perform some task, or even when recruiting agencies match curricula and job profiles. In such open environments, the objective of a matchmaking process is to discover best available offers to a given request. We address the problem of matchmaking from a knowledge representation perspective, with a formalization based on Description Logics. We devise Concept Abduction and Concept Contraction as non-monotonic inferences in Description Logics suitable for modeling matchmaking in a logical framework, and prove some related complexity results. We also present reasonable algorithms for semantic matchmaking based on the devised inferences, and prove that they obey to some commonsense properties. Finally, we report on the implementation of the proposed matchmaking framework, which has been used both as a mediator in e-marketplaces and for semantic web services discovery.

cs.AI↗

arXiv2011

Consistency and Random Constraint Satisfaction Models

In this paper, we study the possibility of designing non-trivial random CSP models by exploiting the intrinsic connection between structures and typical-case hardness. We show that constraint consistency, a notion that has been developed to improve the efficiency of CSP algorithms, is in fact the key to the design of random CSP models that have interesting phase transition behavior and guaranteed exponential resolution complexity without putting much restriction on the parameter of constraint tightness or the domain size of the problem. We propose a very flexible framework for constructing problem instances withinteresting behavior and develop a variety of concrete methods to construct specific random CSP models that enforce different levels of constraint consistency. A series of experimental studies with interesting observations are carried out to illustrate the effectiveness of introducing structural elements in random instances, to verify the robustness of our proposal, and to investigate features of some specific models based on our framework that are highly related to the behavior of backtracking search algorithms.

cs.AI↗

arXiv2011

Solution-Guided Multi-Point Constructive Search for Job Shop Scheduling

Solution-Guided Multi-Point Constructive Search (SGMPCS) is a novel constructive search technique that performs a series of resource-limited tree searches where each search begins either from an empty solution (as in randomized restart) or from a solution that has been encountered during the search. A small number of these "elite solutions is maintained during the search. We introduce the technique and perform three sets of experiments on the job shop scheduling problem. First, a systematic, fully crossed study of SGMPCS is carried out to evaluate the performance impact of various parameter settings. Second, we inquire into the diversity of the elite solution set, showing, contrary to expectations, that a less diverse set leads to stronger performance. Finally, we compare the best parameter setting of SGMPCS from the first two experiments to chronological backtracking, limited discrepancy search, randomized restart, and a sophisticated tabu search algorithm on a set of well-known benchmark problems. Results demonstrate that SGMPCS is significantly better than the other constructive techniques tested, though lags behind the tabu search.

cs.AI↗

arXiv2011

Answer Sets for Logic Programs with Arbitrary Abstract Constraint Atoms

In this paper, we present two alternative approaches to defining answer sets for logic programs with arbitrary types of abstract constraint atoms (c-atoms). These approaches generalize the fixpoint-based and the level mapping based answer set semantics of normal logic programs to the case of logic programs with arbitrary types of c-atoms. The results are four different answer set definitions which are equivalent when applied to normal logic programs. The standard fixpoint-based semantics of logic programs is generalized in two directions, called answer set by reduct and answer set by complement. These definitions, which differ from each other in the treatment of negation-as-failure (naf) atoms, make use of an immediate consequence operator to perform answer set checking, whose definition relies on the notion of conditional satisfaction of c-atoms w.r.t. a pair of interpretations. The other two definitions, called strongly and weakly well-supported models, are generalizations of the notion of well-supported models of normal logic programs to the case of programs with c-atoms. As for the case of fixpoint-based semantics, the difference between these two definitions is rooted in the treatment of naf atoms. We prove that answer sets by reduct (resp. by complement) are equivalent to weakly (resp. strongly) well-supported models of a program, thus generalizing the theorem on the correspondence between stable models and well-supported models of a normal logic program to the class of programs with c-atoms. We show that the newly defined semantics coincide with previously introduced semantics for logic programs with monotone c-atoms, and they extend the original answer set semantics of normal logic programs. We also study some properties of answer sets of programs with c-atoms, and relate our definitions to several semantics for logic programs with aggregates presented in the literature.

cs.AI↗

arXiv2011

The Generalized A* Architecture

We consider the problem of computing a lightest derivation of a global structure using a set of weighted rules. A large variety of inference problems in AI can be formulated in this framework. We generalize A* search and heuristics derived from abstractions to a broad class of lightest derivation problems. We also describe a new algorithm that searches for lightest derivations using a hierarchy of abstractions. Our generalization of A* gives a new algorithm for searching AND/OR graphs in a bottom-up fashion. We discuss how the algorithms described here provide a general architecture for addressing the pipeline problem --- the problem of passing information back and forth between various stages of processing in a perceptual system. We consider examples in computer vision and natural language processing. We apply the hierarchical search algorithm to the problem of estimating the boundaries of convex objects in grayscale images and compare it to other search methods. A second set of experiments demonstrate the use of a new compositional model for finding salient curves in images.

cs.AI↗

arXiv2026

LLM Self-Correction with DeCRIM: Decompose, Critique, and Refine for Enhanced Following of Instructions with Multiple Constraints

Instruction following is a key capability for LLMs. However, recent studies have shown that LLMs often struggle with instructions containing multiple constraints (e.g. a request to create a social media post "in a funny tone" with "no hashtag"). Despite this, most evaluations focus solely on synthetic data. To address this, we introduce RealInstruct, the first benchmark designed to evaluate LLMs' ability to follow real-world multi-constrained instructions by leveraging queries real users asked AI assistants. We also investigate model-based evaluation as a cost-effective alternative to human annotation for this task. Our findings reveal that even the proprietary GPT-4 model fails to meet at least one constraint on over 21% of instructions, highlighting the limitations of state-of-the-art models. To address the performance gap between open-source and proprietary models, we propose the Decompose, Critique and Refine (DeCRIM) self-correction pipeline, which enhances LLMs' ability to follow constraints. DeCRIM works by decomposing the original instruction into a list of constraints and using a Critic model to decide when and where the LLM's response needs refinement. Our results show that DeCRIM improves Mistral's performance by 7.3% on RealInstruct and 8.0% on IFEval even with weak feedback. Moreover, we demonstrate that with strong feedback, open-source LLMs with DeCRIM can outperform GPT-4 on both benchmarks.

cs.CL↗

arXiv2026

GraphDx: A Cost-Aware Knowledge-Enhanced Multi-Agent Framework for Sequential Diagnosis

Sequential diagnosis requires balancing diagnostic accuracy against resource costs through iterative information gathering. Existing Large Language Model (LLM) approaches exhibit a critical knowledge-reasoning gap: despite encoding extensive medical knowledge, they struggle to reason systematically under cost constraints, often resorting to excessive testing. We propose GraphDx, a knowledge-enhanced framework with two core innovations. First, we design an automated pipeline that leverages LLMs to construct Medical Diagnosis Knowledge Graphs (MDKGs) with quantized typicality, action-centric topology, and dual-objective attributes for both diagnostic relevance and cost-sensitivity. Second, we introduce three collaborative agents (Perception, Reasoning, and Decision) where the Perception and Decision Agents handle language understanding and generation, while the Reasoning Agent performs deterministic evidence scoring and cost-aware planning on the MDKG. Experiments on MedQA and MIMIC-IV across three LLM backbones (DeepSeek-V3, Kimi-k2, Llama-3.3) show that GraphDx improves diagnostic success rates from 50--68% to 79--93% while reducing test costs by 20--54%, providing a robust, economical, and interpretable solution for automated clinical diagnosis.

cs.AI↗

arXiv2026

Concept Tokens: Learning Behavioral Embeddings Through Concept Definitions

We propose Concept Tokens, a lightweight method that adds a new special token to a pretrained LLM and learns only its embedding from multiple natural language definitions of a target concept, where occurrences of the concept are replaced by the new token. The LLM is kept frozen and the embedding is optimized with the standard language-modeling objective. We evaluate Concept Tokens in three settings. First, we study hallucinations in closed-book question answering on HotpotQA and find a directional effect: negating the hallucination token reduces hallucinated answers mainly by increasing abstentions, whereas asserting it increases hallucinations and lowers precision. Second, we induce recasting, a pedagogical feedback strategy for second language teaching, and observe the same directional effect. Moreover, compared to providing the full definitional corpus in-context, concept tokens better preserve compliance with other instructions (e.g., asking follow-up questions). Finally, we include a qualitative study with the Eiffel Tower and a fictional "Austral Tower" to illustrate what information the learned embeddings capture and where their limitations emerge. Overall, Concept Tokens provide a compact control signal learned from definitions that can steer behavior in frozen LLMs.

cs.CL↗

arXiv2026

Beyond Single Plots: A Benchmark for Question Answering on Multi-Charts

Charts are widely used to present complex information. Deriving meaningful insights in real-world contexts often requires interpreting multiple related charts together. Research on understanding multi-chart images has not been extensively explored. We introduce PolyChartQA, a mid-scale dataset specifically designed for question answering over multi-chart images. PolyChartQA comprises 534 multi-chart images (with a total of 2,297 sub-charts) sourced from peer-reviewed computer science research publications and 2,694 QA pairs. We evaluate the performance of nine state-of-the-art Multimodal Language Models (MLMs) on PolyChartQA across question type, difficulty, question source, and key structural characteristics of multi-charts. Our results show a 27.4% LLM-based accuracy (L-Accuracy) drop on human-authored questions compared to MLM-generated questions, and a 5.39% L-accuracy gain with our proposed prompting method.

cs.CL↗

arXiv2026

PrinciplismQA: A Philosophy-Grounded Approach to Assessing LLM-Human Clinical Medical Ethics Alignment

As medical LLMs transition to clinical deployment, assessing their ethical reasoning capability becomes critical. While achieving high accuracy on knowledge benchmarks, LLMs lack validated assessment for navigating ethical trade-offs in clinical decision-making where multiple valid solutions exist. Existing benchmarks lack systematic approaches to incorporate recognized philosophical frameworks and expert validation for ethical reasoning assessment. We introduce PrinciplismQA, a philosophy-grounded approach to assessing LLM clinical medical ethics alignment. Grounded in Principlism, our approach provides a systematic methodology for incorporating clinical ethics philosophy into LLM assessment design. PrinciplismQA comprises 3,648 expert-validated questions spanning knowledge assessment and clinical reasoning. Our expert-calibrated pipeline enables reproducible evaluation and models ethical biases. Evaluating recent models reveals significant ethical reasoning gaps despite high knowledge accuracy, demonstrating that knowledge-oriented training does not ensure clinical ethical alignment. PrinciplismQA provides a validated tool for assessing clinical AI deployment readiness. Our data and test scripts are fully released on https://github.com/FreedomIntelligence/PrinciplismQA.

cs.CL↗

arXiv2026

DICA: Dual-Indicator Guided Contrastive Alignment in Multimodal Large Language Models

Human visual reasoning typically follows a coarse-to-fine attention process, starting from global scene understanding and gradually focusing on question-relevant regions. However, multimodal large language models may deviate from this pattern due to attention drift and the underutilization of visual evidence, which can lead to hallucinations. To mitigate these issues, this study proposes a Dual-Indicator Guided Contrastive Alignment (DICA), which tracks two information-theoretic indicators during inference: Visual Attention Entropy (VAE), which reflects the concentration of visual attention, and Output Image Correlation (OIC), which measures the dependence of generated outputs on the visual input. An abnormal increase in VAE or a decrease in OIC corresponds to different failure modes, which trigger targeted contrastive alignment to restore visual grounding. Experimental results across multiple benchmarks demonstrate that DICA consistently outperforms existing approaches and substantially reduces hallucinations, highlighting the effectiveness of indicator-driven intervention in improving multimodal inference reliability. The code is publicly available at https://github.com/BGWH123/DICA/.

cs.AI↗

arXiv2026

A Mechanistic Perspective and Circuit-Guided Difficulty Metric for Unlearning

Machine unlearning is becoming essential for building trustworthy and compliant language models. Yet unlearning success varies considerably across individual samples: some are reliably erased, while others persist despite the same procedure. We argue that this disparity is not only a data-side phenomenon, but also reflects model-internal mechanisms that encode and protect memorized information. We study this problem from a mechanistic perspective based on model circuits--structured interaction pathways that govern how predictions are formed. We propose Circuit-guided Unlearning Difficulty (CUD), a {\em pre-unlearning} metric that assigns each sample a continuous difficulty score using circuit-level signals. Extensive experiments demonstrate that CUD reliably separates intrinsically easy and hard samples, and remains stable across unlearning methods. We identify key circuit-level patterns that reveal a mechanistic signature of difficulty: easy-to-unlearn samples are associated with shorter, shallower interactions concentrated in earlier-to-intermediate parts of the original model, whereas hard samples rely on longer and deeper pathways closer to late-stage computation. Compared to existing qualitative studies, CUD takes a first step toward a principled, fine-grained, and interpretable analysis of unlearning difficulty; and motivates the development of unlearning methods grounded in model mechanisms.

cs.LG↗

arXiv2026

Position: Evaluation Scores Are Perishable Knowledge Claims

Evaluation methodologies for language models increasingly combine multiple signals, from automated metrics and LLM-as-judge ratings to human assessments and benchmark suite results. When these signals are aggregated via averaging, evaluation confidence can then substantially exceed the reliability of the weakest signal: a phenomenon we call trust inflation in evaluation. We argue that evaluation scores should be treated as epistemic claims with three properties: formality (human evaluation provides stronger evidence than an automated metric), scope (a benchmark result applies to the tested distribution, not universally), and validity windows (benchmark results expire as contamination accumulates and distributions shift). Several converging research traditions (chain-of-thought analysis, possibilistic logic, and algebraic theory) establish weakest-link aggregation as the conservative endpoint of a parameterized operator family controlled by a single pessimism parameter. Drawing on those traditions, and on concrete lessons from building an evaluation harness for agentic AI, we propose that evaluation results carry explicit metadata (formality tier, scope declaration, and expiration date) to make their epistemic status transparent. We illustrate the cost of mean aggregation on the public HELM leaderboard: across 54 frontier models on ten scenarios, the top-five models ranked by mean score and by weakest-link are completely disjoint.

cs.AI↗

arXiv2025

Leveraging Fine-Tuned Large Language Models for Interpretable Pancreatic Cystic Lesion Feature Extraction and Risk Categorization

Background: Manual extraction of pancreatic cystic lesion (PCL) features from radiology reports is labor-intensive, limiting large-scale studies needed to advance PCL research. Purpose: To develop and evaluate large language models (LLMs) that automatically extract PCL features from MRI/CT reports and assign risk categories based on guidelines. Materials and Methods: We curated a training dataset of 6,000 abdominal MRI/CT reports (2005-2024) from 5,134 patients that described PCLs. Labels were generated by GPT-4o using chain-of-thought (CoT) prompting to extract PCL and main pancreatic duct features. Two open-source LLMs were fine-tuned using QLoRA on GPT-4o-generated CoT data. Features were mapped to risk categories per institutional guideline based on the 2017 ACR White Paper. Evaluation was performed on 285 held-out human-annotated reports. Model outputs for 100 cases were independently reviewed by three radiologists. Feature extraction was evaluated using exact match accuracy, risk categorization with macro-averaged F1 score, and radiologist-model agreement with Fleiss' Kappa. Results: CoT fine-tuning improved feature extraction accuracy for LLaMA (80% to 97%) and DeepSeek (79% to 98%), matching GPT-4o (97%). Risk categorization F1 scores also improved (LLaMA: 0.95; DeepSeek: 0.94), closely matching GPT-4o (0.97), with no statistically significant differences. Radiologist inter-reader agreement was high (Fleiss' Kappa = 0.888) and showed no statistically significant difference with the addition of DeepSeek-FT-CoT (Fleiss' Kappa = 0.893) or GPT-CoT (Fleiss' Kappa = 0.897), indicating that both models achieved agreement levels on par with radiologists. Conclusion: Fine-tuned open-source LLMs with CoT supervision enable accurate, interpretable, and efficient phenotyping for large-scale PCL research, achieving performance comparable to GPT-4o.

cs.AI↗

arXiv2026

Extending Desbordante with Probabilistic Functional Dependency Discovery Support

Data profiling aims to extract complex patterns from data for further analysis and use that data in domains such as data cleaning, data deduplication, anomaly detection, and many more. Functional dependencies (FDs) are one of the most well-known patterns. However, they are poorly suited for these tasks, as real data is usually dirty, and the rigid definition of FDs does not allow algorithms to locate them. For this reason, there are several formulations aimed at relaxing FDs to support dirty data, with approximate functional dependency (AFD) being the most popular one. Another formulation is the Probabilistic Functional Dependency (pFD), which we aim to support inside Desbordante - a science-intensive, high-performance and open-source data profiling tool implemented in C++. However, pFDs are relatively poorly studied, compared to AFDs. In this paper we study pFDs, both analytically and empirically. We start by assessing how different pFDs and AFDs are by studying cases in which pFDs have an edge over AFDs. Then, we implement the algorithm for pFD discovery, as well as study its run time and memory consumption. We also compare it with an AFD discovery algorithm. Lastly, we study the output of both algorithms to learn whether or not it is possible to use AFD discovery algorithm to get pFDs and vice versa.

cs.DB↗

arXiv2026

A Systems Engineering Framework for Vision-Language-Enabled UAV Triage and Disaster Response

Recent advances in Vision Language Models (VLMs) have created new opportunities for disaster response, where responders must interpret large volumes of sensor data under time pressure. Current VLM applications include social media monitoring for situational awareness, generation of draft action plans, and translation of technical alerts into public-facing messages. While these efforts can accelerate information flow, they remain largely limited to decision-support roles. Such approaches can increase operator burden because humans must still translate outputs into coordinated actions across teams and robotic assets. This study explores the viability of embedding VLMs as coordination agents within the human-UAV loop. The proposed architecture integrates natural language interaction, mission-level task coordination, software-in-the-loop implementation, and communication aligned with the Incident Command System (ICS). Rather than functioning solely as advisory tools, VLMs facilitate communication between human operators, mission control logic, and UAV task execution. The framework was developed using a Model-Based Systems Engineering (MBSE) approach, with use case and block definition diagrams representing system roles, internal structure, and component interactions. Three key elements, the VLM Coordinator Agent, UAV Mission Control, and Task Allocator, were implemented within an integrated simulation and control environment. A preliminary human-factors evaluation with seven participants showed reduced perceived workload across mental demand, effort, and frustration, along with high ratings for AI trust and communication clarity. By integrating MBSE, software-in-the-loop testing, and human-factors evaluation, this work advances scalable human-autonomy teaming for high-stakes disaster response, with broader implications for aerospace autonomy and civil safety.

cs.RO↗

arXiv2026

Building Large-Scale English-Romanian Literary Translation Resources with Open Models

Literary translation has recently gained attention as a distinct and complex task in machine translation research, yet translation by small open models remains an open problem, particularly for low-resource languages such as Romanian. We introduce the TinyFabulist Translation Framework (TF2), a unified framework for dataset creation, fine-tuning, and evaluation in English $\to$ Romanian literary translation. Building on DS-TF1-EN-3M, the largest collection of synthetic English fables to date, our pipeline first generates 15k high-quality Romanian references from the TF1 pool using a high-performing large language model (LLM). We then apply a two-stage fine-tuning process to a 12B-parameter open-weight model: (i) instruction tuning to capture genre-specific narrative style, and (ii) adapter compression for efficient deployment. Evaluation combines a five-dimension LLM-based rubric (accuracy, fluency, coherence, style, cultural adaptation) as the primary comparative framework, alongside corpus-level Bilingual Evaluation Understudy (BLEU) reported as a secondary reference-based consistency metric. Our fine-tuned model (TF2-12B) achieves strong fluency and adequacy, narrowing the gap to top-performing proprietary models under automated and human-anchored evaluation, while being open, accessible, and significantly more cost-effective. We publicly release the fine-tuned model and two large-scale synthetic parallel datasets (DS-TF2-EN-RO-3M and DS-TF2-EN-RO-15K), along with all scripts and evaluation prompts. TF2 provides an end-to-end, reproducible pipeline for research on cost-efficient translation, cross-lingual narrative generation, and the broad adoption of open models for culturally significant literary content in low-resource settings.

cs.CL↗