Tootfinder

Quantifying the Accuracy-Interpretability Trade-Off in Concept-Based Sidechannel Models
David Debot, Giuseppe Marra
https://arxiv.org/abs/2510.05670 https://

Quantifying the Accuracy-Interpretability Trade-Off in Concept-Based Sidechannel Models
Concept Bottleneck Models (CBNMs) are deep learning models that provide interpretability by enforcing a bottleneck layer where predictions are based exclusively on human-understandable concepts. However, this constraint also restricts information flow and often results in reduced predictive accuracy. Concept Sidechannel Models (CSMs) address this limitation by introducing a sidechannel that bypasses the bottleneck and carry additional task-relevant information. While this improves accuracy, it …

DeGuV: Depth-Guided Visual Reinforcement Learning for Generalization and Interpretability in Manipulation
Tien Pham, Xinyun Chi, Khang Nguyen, Manfred Huber, Angelo Cangelosi
https://arxiv.org/abs/2509.04970

DeGuV: Depth-Guided Visual Reinforcement Learning for Generalization and Interpretability in Manipulation
Reinforcement learning (RL) agents can learn to solve complex tasks from visual inputs, but generalizing these learned skills to new environments remains a major challenge in RL application, especially robotics. While data augmentation can improve generalization, it often compromises sample efficiency and training stability. This paper introduces DeGuV, an RL framework that enhances both generalization and sample efficiency. In specific, we leverage a learnable masker network that produces a ma…

Sparse-Group Factor Analysis for High-Dimensional Time Series
Xin Wang, Xialu Liu
https://arxiv.org/abs/2510.05370 https://arxiv.org/pdf/2510.05370

Sparse-Group Factor Analysis for High-Dimensional Time Series
Factor analysis is a widely used technique for dimension reduction in high-dimensional data. However, a key challenge in factor models lies in the interpretability of the latent factors. One intuitive way to interpret these factors is through their associated loadings. Liu and Wang proposed a novel framework that redefines factor models with sparse loadings to enhance interpretability. In many high-dimensional time series applications, variables exhibit natural group structures. Building on thi…

Visual Representations inside the Language Model
Benlin Liu, Amita Kamath, Madeleine Grunde-McLaughlin, Winson Han, Ranjay Krishna
https://arxiv.org/abs/2510.04819 https://

Visual Representations inside the Language Model
Despite interpretability work analyzing VIT encoders and transformer activations, we don't yet understand why Multimodal Language Models (MLMs) struggle on perception-heavy tasks. We offer an under-studied perspective by examining how popular MLMs (LLaVA-OneVision, Qwen2.5-VL, and Llama-3-LLaVA-NeXT) process their visual key-value tokens. We first study the flow of visual information through the language model, finding that image value tokens encode sufficient information to perform several per…

Mechanistic Interpretability of Code Correctness in LLMs via Sparse Autoencoders
Kriz Tahimic, Charibeth Cheng
https://arxiv.org/abs/2510.02917 https://arx…

Mechanistic Interpretability of Code Correctness in LLMs via Sparse Autoencoders
As Large Language Models become integral to software development, with substantial portions of AI-suggested code entering production, understanding their internal correctness mechanisms becomes critical for safe deployment. We apply sparse autoencoders to decompose LLM representations, identifying directions that correspond to code correctness. We select predictor directions using t-statistics and steering directions through separation scores from base model representations, then analyze their …

Replaced article(s) found for cs.AI. https://arxiv.org/list/cs.AI/new
[4/6]:
- Cross-Document Cross-Lingual NLI via RST-Enhanced Graph Fusion and Interpretability Prediction
Mengying Yuan, Wenhao Wang, Zixuan Wang, Yujie Huang, Kangli Wei, Fei Li, Chong Teng, Donghong Ji

CWEFS: Brain volume conduction effects inspired channel-wise EEG feature selection for multi-dimensional emotion recognition
Xueyuan Xu, Wenjia Dong, Fulin Wei, Li Zhuo
https://arxiv.org/abs/2508.05228

CWEFS: Brain volume conduction effects inspired channel-wise EEG feature selection for multi-dimensional emotion recognition
Due to the intracranial volume conduction effects, high-dimensional multi-channel electroencephalography (EEG) features often contain substantial redundant and irrelevant information. This issue not only hinders the extraction of discriminative emotional representations but also compromises the real-time performance. Feature selection has been established as an effective approach to address the challenges while enhancing the transparency and interpretability of emotion recognition models. Howev…

Optimal Regularization Under Uncertainty: Distributional Robustness and Convexity Constraints
Oscar Leong, Eliza O'Reilly, Yong Sheng Soh
https://arxiv.org/abs/2510.03464 ht…

Optimal Regularization Under Uncertainty: Distributional Robustness and Convexity Constraints
Regularization is a central tool for addressing ill-posedness in inverse problems and statistical estimation, with the choice of a suitable penalty often determining the reliability and interpretability of downstream solutions. While recent work has characterized optimal regularizers for well-specified data distributions, practical deployments are often complicated by distributional uncertainty and the need to enforce structural constraints such as convexity. In this paper, we introduce a frame…

D\'esentrelacement Fr\'equentiel Doux pour les Codecs Audio Neuronaux
Beno\^it Gini\`es, Xiaoyu Bie, Olivier Fercoq, Ga\"el Richard
https://arxiv.org/abs/2510.03741

Désentrelacement Fréquentiel Doux pour les Codecs Audio Neuronaux
While neural-based models have led to significant advancements in audio feature extraction, the interpretability of the learned representations remains a critical challenge. To address this, disentanglement techniques have been integrated into discrete neural audio codecs to impose structure on the extracted tokens. However, these approaches often exhibit strong dependencies on specific datasets or task formulations. In this work, we propose a disentangled neural audio codec that leverages spec…

KGRAG-SC: Knowledge Graph RAG-Assisted Semantic Communication
Dayu Fan, Rui Meng, Song Gao, Xiaodong Xu
https://arxiv.org/abs/2509.04801 https://arxiv.org/…

KGRAG-SC: Knowledge Graph RAG-Assisted Semantic Communication
The state-of-the-art semantic communication (SC) schemes typically rely on end-to-end deep learning frameworks that lack interpretability and struggle with robust semantic selection and reconstruction under noisy conditions. To address this issue, this paper presents KGRAG-SC, a knowledge graph-assisted SC framework that leverages retrieval-augmented generation principles. KGRAG-SC employs a multi-dimensional knowledge graph, enabling efficient semantic extraction through community-guided entit…

Interpreting anomaly detection of SDSS spectra
Edgar Ortiz Manrique, M\'ed\'eric Boquien
https://arxiv.org/abs/2510.05235 https://arxiv.org/pdf/251…

Interpreting anomaly detection of SDSS spectra
The increasing use of ML in astronomy introduces important questions about interpretability. Due to their complexity and non-linear nature, it can be challenging to understand their decision-making process. While these models can effectively identify unusual spectra, interpreting the physical nature of the flagged outliers remains a major challenge. We aim to bridge the gap between anomaly detection and physical understanding by combining deep learning with interpretable ML (iML) techniques to …

Enhancing Interpretability and Effectiveness in Recommendation with Numerical Features via Learning to Contrast the Counterfactual samples
Xiaoxiao Xu, Hao Wu, Wenhui Yu, Lantao Hu, Peng Jiang, Kun Gai
https://arxiv.org/abs/2509.03187

Enhancing Interpretability and Effectiveness in Recommendation with Numerical Features via Learning to Contrast the Counterfactual samples
We propose a general model-agnostic Contrastive learning framework with Counterfactual Samples Synthesizing (CCSS) for modeling the monotonicity between the neural network output and numerical features which is critical for interpretability and effectiveness of recommender systems. CCSS models the monotonicity via a two-stage process: synthesizing counterfactual samples and contrasting the counterfactual samples. The two techniques are naturally integrated into a model-agnostic framework, formi…

Atlas-free Brain Network Transformer
Shuai Huang, Xuan Kan, James J. Lah, Deqiang Qiu
https://arxiv.org/abs/2510.03306 https://arxiv.org/pdf/2510.03306

Atlas-free Brain Network Transformer
Current atlas-based approaches to brain network analysis rely heavily on standardized anatomical or connectivity-driven brain atlases. However, these fixed atlases often introduce significant limitations, such as spatial misalignment across individuals, functional heterogeneity within predefined regions, and atlas-selection biases, collectively undermining the reliability and interpretability of the derived brain networks. To address these challenges, we propose a novel atlas-free brain network…

Identifying Exoplanets with Deep Learning: A CNN and RNN Classifier for Kepler DR25 and Candidate Vetting
Bibin Thomas, Vittal Bhat M, Salman Arafath Mohammed, Abdul Wase Mohammed, Adis Abebaw Dessalegn, Mohit Mittal
https://arxiv.org/abs/2509.04793

Identifying Exoplanets with Deep Learning: A CNN and RNN Classifier for Kepler DR25 and Candidate Vetting
The rapid expansion of exoplanet survey missions such as Kepler, TESS, and the upcoming PLATO mission has generated massive light-curve datasets that challenge traditional vetting pipelines. We introduce a hybrid deep-learning framework that integrates convolutional networks, bidirectional LSTMs, and an attention mechanism to identify planetary transit signals with improved accuracy and interpretability. Trained on Kepler DR25 data, the model achieves F1 = $0.910 \pm 0.008$ (AUC--ROC = $0.984 \…

TopInG: Topologically Interpretable Graph Learning via Persistent Rationale Filtration
Cheng Xin, Fan Xu, Xin Ding, Jie Gao, Jiaxin Ding
https://arxiv.org/abs/2510.05102 https:/…

TopInG: Topologically Interpretable Graph Learning via Persistent Rationale Filtration
Graph Neural Networks (GNNs) have shown remarkable success across various scientific fields, yet their adoption in critical decision-making is often hindered by a lack of interpretability. Recently, intrinsically interpretable GNNs have been studied to provide insights into model predictions by identifying rationale substructures in graphs. However, existing methods face challenges when the underlying rationale subgraphs are complex and varied. In this work, we propose TopInG: Topologically Int…

High-Resolution Global Land Surface Temperature Retrieval via a Coupled Mechanism-Machine Learning Framework
Tian Xie, Huanfeng Shen, Menghui Jiang, Juan-Carlos Jim\'enez-Mu\~noz, Jos\'e A. Sobrino, Huifang Li, Chao Zeng
https://arxiv.org/abs/2509.04991

High-Resolution Global Land Surface Temperature Retrieval via a Coupled Mechanism-Machine Learning Framework
Land surface temperature (LST) is vital for land-atmosphere interactions and climate processes. Accurate LST retrieval remains challenging under heterogeneous land cover and extreme atmospheric conditions. Traditional split window (SW) algorithms show biases in humid environments; purely machine learning (ML) methods lack interpretability and generalize poorly with limited data. We propose a coupled mechanism model-ML (MM-ML) framework integrating physical constraints with data-driven learning …

Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning?
Qingyu Yin, Chak Tou Leong, Linyi Yang, Wenxuan Huang, Wenjie Li, Xiting Wang, Jaehong Yoon, YunXing, XingYu, Jinjin Gu
https://arxiv.org/abs/2510.06036

Refusal Falls off a Cliff: How Safety Alignment Fails in Reasoning?
Large reasoning models (LRMs) with multi-step reasoning capabilities have shown remarkable problem-solving abilities, yet they exhibit concerning safety vulnerabilities that remain poorly understood. In this work, we investigate why safety alignment fails in reasoning models through a mechanistic interpretability lens. Using a linear probing approach to trace refusal intentions across token positions, we discover a striking phenomenon termed as \textbf{refusal cliff}: many poorly-aligned reason…

Amplitude-based Input Attribution in Quantum Learning via Integrated Gradients
Nicholas S. DiBrita, Jason Han, Younghyun Cho, Hengrui Luo, Tirthak Patel
https://arxiv.org/abs/2510.02497

Amplitude-based Input Attribution in Quantum Learning via Integrated Gradients
Quantum machine learning (QML) algorithms have demonstrated early promise across hardware platforms, but remain difficult to interpret due to the inherent opacity of quantum state evolution. Widely used classical interpretability methods, such as integrated gradients and surrogate-based sensitivity analysis, are not directly compatible with quantum circuits due to measurement collapse and the exponential complexity of simulating state evolution. In this work, we introduce HATTRIQ, a general-pur…

Combining feature-based approaches with graph neural networks and symbolic regression for synergistic performance and interpretability
Rog\'erio Almeida Gouv\^ea, Pierre-Paul De Breuck, Tatiane Pretto, Gian-Marco Rignanese, Marcos Jos\'e Leite dos Santos
https://arxiv.org/abs/2509.03547

Combining feature-based approaches with graph neural networks and symbolic regression for synergistic performance and interpretability
This study introduces MatterVial, an innovative hybrid framework for feature-based machine learning in materials science. MatterVial expands the feature space by integrating latent representations from a diverse suite of pretrained graph neural network (GNN) models including: structure-based (MEGNet), composition-based (ROOST), and equivariant (ORB) graph networks, with computationally efficient, GNN-approximated descriptors and novel features from symbolic regression. Our approach combines the…

EmbodiedCoder: Parameterized Embodied Mobile Manipulation via Modern Coding Model
Zefu Lin, Rongxu Cui, Chen Hanning, Xiangyu Wang, Junjia Xu, Xiaojuan Jin, Chen Wenbo, Hui Zhou, Lue Fan, Wenling Li, Zhaoxiang Zhang
https://arxiv.org/abs/2510.06207

EmbodiedCoder: Parameterized Embodied Mobile Manipulation via Modern Coding Model
Recent advances in control robot methods, from end-to-end vision-language-action frameworks to modular systems with predefined primitives, have advanced robots' ability to follow natural language instructions. Nonetheless, many approaches still struggle to scale to diverse environments, as they often rely on large annotated datasets and offer limited interpretability.In this work, we introduce EmbodiedCoder, a training-free framework for open-world mobile robot manipulation that leverages codin…

Teaching Machines to Speak Using Articulatory Control
Akshay Anand, Chenxu Guo, Cheol Jun Cho, Jiachen Lian, Gopala Anumanchipalli
https://arxiv.org/abs/2510.05619 https://

Teaching Machines to Speak Using Articulatory Control
Current speech production systems predominantly rely on large transformer models that operate as black boxes, providing little interpretability or grounding in the physical mechanisms of human speech. We address this limitation by proposing a new framework: speech generation through explicit articulatory control. This reframes speech as a motor control task similar to robotic manipulation. Our approach uses reinforcement learning to train a policy that directly controls the movements of vocal t…

Exact and Heuristic Algorithms for Constrained Biclustering
Antonio M. Sudoso
https://arxiv.org/abs/2508.05493 https://arxiv.org/pdf/2508.05493

Exact and Heuristic Algorithms for Constrained Biclustering
Biclustering, also known as co-clustering or two-way clustering, simultaneously partitions the rows and columns of a data matrix to reveal submatrices with coherent patterns. Incorporating background knowledge into clustering to enhance solution quality and interpretability has attracted growing interest in mathematical optimization and machine learning research. Extending this paradigm to biclustering enables prior information to guide the joint grouping of rows and columns. We study constrain…

Soft Disentanglement in Frequency Bands for Neural Audio Codecs
Benoit Ginies, Xiaoyu Bie, Olivier Fercoq, Ga\"el Richard
https://arxiv.org/abs/2510.03735 https://

Soft Disentanglement in Frequency Bands for Neural Audio Codecs
In neural-based audio feature extraction, ensuring that representations capture disentangled information is crucial for model interpretability. However, existing disentanglement methods often rely on assumptions that are highly dependent on data characteristics or specific tasks. In this work, we introduce a generalizable approach for learning disentangled features within a neural architecture. Our method applies spectral decomposition to time-domain signals, followed by a multi-branch audio co…

Learning from Failures: Understanding LLM Alignment through Failure-Aware Inverse RL
Nyal Patel, Matthieu Bou, Arjun Jagota, Satyapriya Krishna, Sonali Parbhoo
https://arxiv.org/abs/2510.06092

Learning from Failures: Understanding LLM Alignment through Failure-Aware Inverse RL
Reinforcement Learning from Human Feedback (RLHF) aligns Large Language Models (LLMs) with human preferences, yet the underlying reward signals they internalize remain hidden, posing a critical challenge for interpretability and safety. Existing approaches attempt to extract these latent incentives using Inverse Reinforcement Learning (IRL), but treat all preference pairs equally, often overlooking the most informative signals: those examples the extracted reward model misclassifies or assigns …

Beyond Regularization: Inherently Sparse Principal Component Analysis
Jan O. Bauer
https://arxiv.org/abs/2510.03729 https://arxiv.org/pdf/2510.03729…

Beyond Regularization: Inherently Sparse Principal Component Analysis
Sparse principal component analysis (sparse PCA) is a widely used technique for dimensionality reduction in multivariate analysis, addressing two key limitations of standard PCA. First, sparse PCA can be implemented in high-dimensional low sample size settings, such as genetic microarrays. Second, it improves interpretability as components are regularized to zero. However, over-regularization of sparse singular vectors can cause them to deviate greatly from the population singular vectors, pote…

An Approach to Grounding AI Model Evaluations in Human-derived Criteria
Sasha Mitts
https://arxiv.org/abs/2509.04676 https://arxiv.org/pdf/2509.04676

An Approach to Grounding AI Model Evaluations in Human-derived Criteria
In the rapidly evolving field of artificial intelligence (AI), traditional benchmarks can fall short in attempting to capture the nuanced capabilities of AI models. We focus on the case of physical world modeling and propose a novel approach to augment existing benchmarks with human-derived evaluation criteria, aiming to enhance the interpretability and applicability of model behaviors. Grounding our study in the Perception Test and OpenEQA benchmarks, we conducted in-depth interviews and large…

A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models
Yanbo Wang, Yongcan Yu, Jian Liang, Ran He
https://arxiv.org/abs/2509.03871 https://

A Comprehensive Survey on Trustworthiness in Reasoning with Large Language Models
The development of Long-CoT reasoning has advanced LLM performance across various tasks, including language understanding, complex problem solving, and code generation. This paradigm enables models to generate intermediate reasoning steps, thereby improving both accuracy and interpretability. However, despite these advancements, a comprehensive understanding of how CoT-based reasoning affects the trustworthiness of language models remains underdeveloped. In this paper, we survey recent work on …

SAE-RNA: A Sparse Autoencoder Model for Interpreting RNA Language Model Representations
Taehan Kim, Sangdae Nam
https://arxiv.org/abs/2510.02734 https://ar…

SAE-RNA: A Sparse Autoencoder Model for Interpreting RNA Language Model Representations
Deep learning, particularly with the advancement of Large Language Models, has transformed biomolecular modeling, with protein advances (e.g., ESM) inspiring emerging RNA language models such as RiNALMo. Yet how and what these RNA Language Models internally encode about messenger RNA (mRNA) or non-coding RNA (ncRNA) families remains unclear. We present SAE- RNA, interpretability model that analyzes RiNALMo representations and maps them to known human-level biological features. Our work frames R…

Deep Reinforcement Learning for Ranking Utility Tuning in the Ad Recommender System at Pinterest
Xiao Yang, Mehdi Ben Ayed, Longyu Zhao, Fan Zhou, Yuchen Shen, Abe Engle, Jinfeng Zhuang, Ling Leng, Jiajing Xu, Charles Rosenberg, Prathibha Deshikachar
https://arxiv.org/abs/2509.05292

Deep Reinforcement Learning for Ranking Utility Tuning in the Ad Recommender System at Pinterest
The ranking utility function in an ad recommender system, which linearly combines predictions of various business goals, plays a central role in balancing values across the platform, advertisers, and users. Traditional manual tuning, while offering simplicity and interpretability, often yields suboptimal results due to its unprincipled tuning objectives, the vast amount of parameter combinations, and its lack of personalization and adaptability to seasonality. In this work, we propose a general…

Sparse Deep Additive Model with Interactions: Enhancing Interpretability and Predictability
Yi-Ting Hung, Li-Hsiang Lin, Vince D. Calhoun
https://arxiv.org/abs/2509.23068 https:…

Sparse Deep Additive Model with Interactions: Enhancing Interpretability and Predictability
Recent advances in deep learning highlight the need for personalized models that can learn from small or moderate samples, handle high dimensional features, and remain interpretable. To address this challenge, we propose the Sparse Deep Additive Model with Interactions (SDAMI), a framework that combines sparsity driven feature selection with deep subnetworks for flexible function approximation. Unlike conventional deep learning models, which often function as black boxes, SDAMI explicitly disen…

Statistical Crime Linkage: Evaluating approaches within the Covenant for Using AI in Policing
Nathan A. Judd, Amy V. Tansell, Benjamin Costello, Liam Leonard, Jessica Woodhams, Rowland G. Seymour
https://arxiv.org/abs/2510.03730

Statistical Crime Linkage: Evaluating approaches within the Covenant for Using AI in Policing
Linking crimes by modus operandi has long been employed as an effective tool for crime investigation. The standard statistical method that underpins statistical crime linkage has been logistic regression. The simplicity and interpretability of this approach has been seen as an advantage for law enforcement agencies using statistical crime linkage. In 2023, the National Police Chiefs' Council published the Covenant for Using Artificial Intelligence in Policing designed to guide the development o…

QDeepGR4J: Quantile-based ensemble of deep learning and GR4J hybrid rainfall-runoff models for extreme flow prediction with uncertainty quantification
Arpit Kapoor, Rohitash Chandra
https://arxiv.org/abs/2510.05453

QDeepGR4J: Quantile-based ensemble of deep learning and GR4J hybrid rainfall-runoff models for extreme flow prediction with uncertainty quantification
Conceptual rainfall-runoff models aid hydrologists and climate scientists in modelling streamflow to inform water management practices. Recent advances in deep learning have unravelled the potential for combining hydrological models with deep learning models for better interpretability and improved predictive performance. In our previous work, we introduced DeepGR4J, which enhanced the GR4J conceptual rainfall-runoff model using a deep learning model to serve as a surrogate for the routing comp…

Uncertainty-Aware Concept Bottleneck Models with Enhanced Interpretability
Haifei Zhang, Patrick Barry, Eduardo Brandao
https://arxiv.org/abs/2510.00773 https://

Uncertainty-Aware Concept Bottleneck Models with Enhanced Interpretability
In the context of image classification, Concept Bottleneck Models (CBMs) first embed images into a set of human-understandable concepts, followed by an intrinsically interpretable classifier that predicts labels based on these intermediate representations. While CBMs offer a semantically meaningful and interpretable classification pipeline, they often sacrifice predictive performance compared to end-to-end convolutional neural networks. Moreover, the propagation of uncertainty from concept pred…

Beyond Interpretability: Exploring the Comprehensibility of Adaptive Video Streaming through Large Language Models
Lianchen Jia, Chaoyang Li, Ziqi Yuan, Jiahui Chen, Tianchi Huang, Jiangchuan Liu, Lifeng Sun
https://arxiv.org/abs/2508.16448

Beyond Interpretability: Exploring the Comprehensibility of Adaptive Video Streaming through Large Language Models
Over the past decade, adaptive video streaming technology has witnessed significant advancements, particularly driven by the rapid evolution of deep learning techniques. However, the black-box nature of deep learning algorithms presents challenges for developers in understanding decision-making processes and optimizing for specific application scenarios. Although existing research has enhanced algorithm interpretability through decision tree conversion, interpretability does not directly equate…

On the Optimization of Methods for Establishing Well-Connected Communities
Mohammad Dindoost, Oliver Alvarado Rodriguez, Bartosz Bryg, Minhyuk Park, George Chacko, Tandy Warnow, David A. Bader
https://arxiv.org/abs/2509.02590

On the Optimization of Methods for Establishing Well-Connected Communities
Community detection plays a central role in uncovering meso scale structures in networks. However, existing methods often suffer from disconnected or weakly connected clusters, undermining interpretability and robustness. Well-Connected Clusters (WCC) and Connectivity Modifier (CM) algorithms are post-processing techniques that improve the accuracy of many clustering methods. However, they are computationally prohibitive on massive graphs. In this work, we present optimized parallel implementat…

An Analysis of the New EU AI Act and A Proposed Standardization Framework for Machine Learning Fairness
Mike Teodorescu, Yongxu Sun, Haren N. Bhatia, Christos Makridis
https://arxiv.org/abs/2510.01281 …

An Analysis of the New EU AI Act and A Proposed Standardization Framework for Machine Learning Fairness
The European Union's AI Act represents a crucial step towards regulating ethical and responsible AI systems. However, we find an absence of quantifiable fairness metrics and the ambiguity in terminology, particularly the interchangeable use of the keywords transparency, explainability, and interpretability in the new EU AI Act and no reference of transparency of ethical compliance. We argue that this ambiguity creates substantial liability risk that would deter investment. Fairness transparency…

GFSR-Net: Guided Focus via Segment-Wise Relevance Network for Interpretable Deep Learning in Medical Imaging
Jhonatan Contreras, Thomas Bocklitz
https://arxiv.org/abs/2510.01919

GFSR-Net: Guided Focus via Segment-Wise Relevance Network for Interpretable Deep Learning in Medical Imaging
Deep learning has achieved remarkable success in medical image analysis, however its adoption in clinical practice is limited by a lack of interpretability. These models often make correct predictions without explaining their reasoning. They may also rely on image regions unrelated to the disease or visual cues, such as annotations, that are not present in real-world conditions. This can reduce trust and increase the risk of misleading diagnoses. We introduce the Guided Focus via Segment-Wise R…

Protocode: Prototype-Driven Interpretability for Code Generation in LLMs
Krishna Vamshi Bodla, Haizhao Yang
https://arxiv.org/abs/2509.25247 https://arxiv.…

Protocode: Prototype-Driven Interpretability for Code Generation in LLMs
Since the introduction of Large Language Models (LLMs), they have been widely adopted for various tasks such as text summarization, question answering, speech-to-text translation, and more. In recent times, the use of LLMs for code generation has gained significant attention, with tools such as Cursor and Windsurf demonstrating the ability to analyze massive code repositories and recommend relevant changes. Big tech companies have also acknowledged the growing reliance on LLMs for code generati…

Reduce-Rank Matrix Integer-Valued Autoregressive Model
Kaiyan Cui, Tianyun Guo, Suping Wang
https://arxiv.org/abs/2509.03338 https://arxiv.org/pdf/2509.033…

Reduce-Rank Matrix Integer-Valued Autoregressive Model
Integer-valued time series are widely present in many fields, such as finance, economics, disease transmission, and traffic flow. With data dimensions surging, the traditional multivariate generalized integer autoregressive (MGINAR) model faces parameter overload, poor interpretability, and structural information loss. Matrix integer-valued autoregression (MINAR) model captures row-column cross-correlations and reduces the number of parameters to be estimated. However, further growth in dimensi…

Comparative Field Deployment of Reinforcement Learning and Model Predictive Control for Residential HVAC
Ozan Baris Mulayim, Elias N. Pergantis, Levi D. Reyes Premer, Bingqing Chen, Guannan Qu, Kevin J. Kircher, Mario Berg\'es
https://arxiv.org/abs/2510.01475

Comparative Field Deployment of Reinforcement Learning and Model Predictive Control for Residential HVAC
Advanced control strategies like Model Predictive Control (MPC) offer significant energy savings for HVAC systems but often require substantial engineering effort, limiting scalability. Reinforcement Learning (RL) promises greater automation and adaptability, yet its practical application in real-world residential settings remains largely undemonstrated, facing challenges related to safety, interpretability, and sample efficiency. To investigate these practical issues, we performed a direct com…

SpliDT: Partitioned Decision Trees for Scalable Stateful Inference at Line Rate
Murayyiam Parvez, Annus Zulfiqar, Roman Beltiukov, Shir Landau Feibish, Walter Willinger, Arpit Gupta, Muhammad Shahbaz
https://arxiv.org/abs/2509.00397

SpliDT: Partitioned Decision Trees for Scalable Stateful Inference at Line Rate
Machine learning (ML) is increasingly being deployed in programmable data planes (switches and SmartNICs) to enable real-time traffic analysis, security monitoring, and in-network decision-making. Decision trees (DTs) are particularly well-suited for these tasks due to their interpretability and compatibility with data-plane architectures, i.e., match-action tables (MATs). However, existing in-network DT implementations are constrained by the need to compute all input features upfront, forcing …

Smoothing-Based Conformal Prediction for Balancing Efficiency and Interpretability
Mingyi Zheng, Hongyu Jiang, Yizhou Lu, Jiaye Teng
https://arxiv.org/abs/2509.22529 https://

Smoothing-Based Conformal Prediction for Balancing Efficiency and Interpretability
Conformal Prediction (CP) is a distribution-free framework for constructing statistically rigorous prediction sets. While popular variants such as CD-split improve CP's efficiency, they often yield prediction sets composed of multiple disconnected subintervals, which are difficult to interpret. In this paper, we propose SCD-split, which incorporates smoothing operations into the CP framework. Such smoothing operations potentially help merge the subintervals, thus leading to interpretable predic…

Backdoor Attribution: Elucidating and Controlling Backdoor in Language Models
Miao Yu, Zhenhong Zhou, Moayad Aloqaily, Kun Wang, Biwei Huang, Stephen Wang, Yueming Jin, Qingsong Wen
https://arxiv.org/abs/2509.21761

Backdoor Attribution: Elucidating and Controlling Backdoor in Language Models
Fine-tuned Large Language Models (LLMs) are vulnerable to backdoor attacks through data poisoning, yet the internal mechanisms governing these attacks remain a black box. Previous research on interpretability for LLM safety tends to focus on alignment, jailbreak, and hallucination, but overlooks backdoor mechanisms, making it difficult to understand and fully eliminate the backdoor threat. In this paper, aiming to bridge this gap, we explore the interpretable mechanisms of LLM backdoors through…

Commutative algebra neural network reveals genetic origins of diseases
JunJie Wee, Faisal Suwayyid, Mushal Zia, Hongsong Feng, Yuta Hozumi, Guo-Wei Wei
https://arxiv.org/abs/2509.26566

Commutative algebra neural network reveals genetic origins of diseases
Genetic mutations can disrupt protein structure, stability, and solubility, contributing to a wide range of diseases. Existing predictive models often lack interpretability and fail to integrate physical and chemical interactions critical to molecular mechanisms. Moreover, current approaches treat disease association, stability changes, and solubility alterations as separate tasks, limiting model generalizability. In this study, we introduce a unified framework based on multiscale commutative a…

Efficient Sketching and Nearest Neighbor Search Algorithms for Sparse Vector Sets
Sebastian Bruch, Franco Maria Nardini, Cosimo Rulli, Rossano Venturini
https://arxiv.org/abs/2509.24815

Efficient Sketching and Nearest Neighbor Search Algorithms for Sparse Vector Sets
Sparse embeddings of data form an attractive class due to their inherent interpretability: Every dimension is tied to a term in some vocabulary, making it easy to visually decipher the latent space. Sparsity, however, poses unique challenges for Approximate Nearest Neighbor Search (ANNS) which finds, from a collection of vectors, the k vectors closest to a query. To encourage research on this underexplored topic, sparse ANNS featured prominently in a BigANN Challenge at NeurIPS 2023, where appr…

eDIF: A European Deep Inference Fabric for Remote Interpretability of LLM
Irma Heithoff. Marc Guggenberger, Sandra Kalogiannis, Susanne Mayer, Fabian Maag, Sigurd Schacht, Carsten Lanquillon
https://arxiv.org/abs/2508.10553

eDIF: A European Deep Inference Fabric for Remote Interpretability of LLM
This paper presents a feasibility study on the deployment of a European Deep Inference Fabric (eDIF), an NDIF-compatible infrastructure designed to support mechanistic interpretability research on large language models. The need for widespread accessibility of LLM interpretability infrastructure in Europe drives this initiative to democratize advanced model analysis capabilities for the research community. The project introduces a GPU-based cluster hosted at Ansbach University of Applied Scienc…

The Loss Kernel: A Geometric Probe for Deep Learning Interpretability
Maxwell Adam, Zach Furman, Jesse Hoogland
https://arxiv.org/abs/2509.26537 https://ar…

The Loss Kernel: A Geometric Probe for Deep Learning Interpretability
We introduce the loss kernel, an interpretability method for measuring similarity between data points according to a trained neural network. The kernel is the covariance matrix of per-sample losses computed under a distribution of low-loss-preserving parameter perturbations. We first validate our method on a synthetic multitask problem, showing it separates inputs by task as predicted by theory. We then apply this kernel to Inception-v1 to visualize the structure of ImageNet, and we show that t…

A Foundation Model for Chest X-ray Interpretation with Grounded Reasoning via Online Reinforcement Learning
Qika Lin, Yifan Zhu, Bin Pu, Ling Huang, Haoran Luo, Jingying Ma, Zhen Peng, Tianzhe Zhao, Fangzhi Xu, Jian Zhang, Kai He, Zhonghong Ou, Swapnil Mishra, Mengling Feng
https://arxiv.org/abs/2509.03906

A Foundation Model for Chest X-ray Interpretation with Grounded Reasoning via Online Reinforcement Learning
Medical foundation models (FMs) have shown tremendous promise amid the rapid advancements in artificial intelligence (AI) technologies. However, current medical FMs typically generate answers in a black-box manner, lacking transparent reasoning processes and locally grounded interpretability, which hinders their practical clinical deployments. To this end, we introduce DeepMedix-R1, a holistic medical FM for chest X-ray (CXR) interpretation. It leverages a sequential training pipeline: initiall…

DRetNet: A Novel Deep Learning Framework for Diabetic Retinopathy Diagnosis
Idowu Paul Okuwobi, Jingyuan Liu, Jifeng Wan, Jiaojiao Jiang
https://arxiv.org/abs/2509.01072 https:/…

DRetNet: A Novel Deep Learning Framework for Diabetic Retinopathy Diagnosis
Diabetic retinopathy (DR) is a leading cause of blindness worldwide, necessitating early detection to prevent vision loss. Current automated DR detection systems often struggle with poor-quality images, lack interpretability, and insufficient integration of domain-specific knowledge. To address these challenges, we introduce a novel framework that integrates three innovative contributions: (1) Adaptive Retinal Image Enhancement Using Physics-Informed Neural Networks (PINNs): this technique dyna…

Beyond Transcription: Mechanistic Interpretability in ASR
Neta Glazer, Yael Segal-Feldman, Hilit Segev, Aviv Shamsian, Asaf Buchnick, Gill Hetz, Ethan Fetaya, Joseph Keshet, Aviv Navon
https://arxiv.org/abs/2508.15882

Beyond Transcription: Mechanistic Interpretability in ASR
Interpretability methods have recently gained significant attention, particularly in the context of large language models, enabling insights into linguistic representations, error detection, and model behaviors such as hallucinations and repetitions. However, these techniques remain underexplored in automatic speech recognition (ASR), despite their potential to advance both the performance and interpretability of ASR systems. In this work, we adapt and systematically apply established interpret…

Bayesian Additive Regression Trees for functional ANOVA model
Seokhun Park, Insung Kong, Yongdai Kim
https://arxiv.org/abs/2509.03317 https://arxiv.org/pdf…

Bayesian Additive Regression Trees for functional ANOVA model
Bayesian Additive Regression Trees (BART) is a powerful statistical model that leverages the strengths of Bayesian inference and regression trees. It has received significant attention for capturing complex non-linear relationships and interactions among predictors. However, the accuracy of BART often comes at the cost of interpretability. To address this limitation, we propose ANOVA Bayesian Additive Regression Trees (ANOVA-BART), a novel extension of BART based on the functional ANOVA decompo…

AutoDrive-R$^2$: Incentivizing Reasoning and Self-Reflection Capacity for VLA Model in Autonomous Driving
Zhenlong Yuan, Jing Tang, Jinguo Luo, Rui Chen, Chengxuan Qian, Lei Sun, Xiangxiang Chu, Yujun Cai, Dapeng Zhang, Shuo Li
https://arxiv.org/abs/2509.01944

AutoDrive-R$^2$: Incentivizing Reasoning and Self-Reflection Capacity for VLA Model in Autonomous Driving
Vision-Language-Action (VLA) models in autonomous driving systems have recently demonstrated transformative potential by integrating multimodal perception with decision-making capabilities. However, the interpretability and coherence of the decision process and the plausibility of action sequences remain largely underexplored. To address these issues, we propose AutoDrive-R$^2$, a novel VLA framework that enhances both reasoning and self-reflection capabilities of autonomous driving systems thr…

Assessing the Noise Robustness of Class Activation Maps: A Framework for Reliable Model Interpretability
Syamantak Sarkar, Revoti P. Bora, Bhupender Kaushal, Sudhish N George, Kiran Raja
https://arxiv.org/abs/2508.18154

Assessing the Noise Robustness of Class Activation Maps: A Framework for Reliable Model Interpretability
Class Activation Maps (CAMs) are one of the important methods for visualizing regions used by deep learning models. Yet their robustness to different noise remains underexplored. In this work, we evaluate and report the resilience of various CAM methods for different noise perturbations across multiple architectures and datasets. By analyzing the influence of different noise types on CAM explanations, we assess the susceptibility to noise and the extent to which dataset characteristics may impa…

Interpreting Language Models Through Concept Descriptions: A Survey
Nils Feldhus, Laura Kopf
https://arxiv.org/abs/2510.01048 https://arxiv.org/pdf/2510.01…

Interpreting Language Models Through Concept Descriptions: A Survey
Understanding the decision-making processes of neural networks is a central goal of mechanistic interpretability. In the context of Large Language Models (LLMs), this involves uncovering the underlying mechanisms and identifying the roles of individual model components such as neurons and attention heads, as well as model abstractions such as the learned sparse features extracted by Sparse Autoencoders (SAEs). A rapidly growing line of work tackles this challenge by using powerful generator mod…

Open Opportunities in AI Safety, Alignment, and Ethics (AI SAE)
Dylan Waldner
https://arxiv.org/abs/2509.24065 https://arxiv.org/pdf/2509.24065

Open Opportunities in AI Safety, Alignment, and Ethics (AI SAE)
AI safety research has emphasized interpretability, control, and robustness, yet without an ethical substrate these approaches may remain fragile under competitive and open-ended pressures. This paper explores ethics not as an external add-on, but as a possible structural lens for alignment, introducing a \emph{moral problem space} $M$: a high-dimensional domain in which moral distinctions could, in principle, be represented in AI systems. Human moral reasoning is treated as a compressed and su…

Machine Intelligence on the Edge: Interpretable Cardiac Pattern Localisation Using Reinforcement Learning
Haozhe Tian, Qiyu Rao, Nina Moutonnet, Pietro Ferraro, Danilo Mandic
https://arxiv.org/abs/2508.21652

Machine Intelligence on the Edge: Interpretable Cardiac Pattern Localisation Using Reinforcement Learning
Matched filters are widely used to localise signal patterns due to their high efficiency and interpretability. However, their effectiveness deteriorates for low signal-to-noise ratio (SNR) signals, such as those recorded on edge devices, where prominent noise patterns can closely resemble the target within the limited length of the filter. One example is the ear-electrocardiogram (ear-ECG), where the cardiac signal is attenuated and heavily corrupted by artefacts. To address this, we propose th…

Analyzing Latent Concepts in Code Language Models
Arushi Sharma, Vedant Pungliya, Christopher J. Quinn, Ali Jannesari
https://arxiv.org/abs/2510.00476 https://

Analyzing Latent Concepts in Code Language Models
Interpreting the internal behavior of large language models trained on code remains a critical challenge, particularly for applications demanding trust, transparency, and semantic robustness. We propose Code Concept Analysis (CoCoA): a global post-hoc interpretability framework that uncovers emergent lexical, syntactic, and semantic structures in a code language model's representation space by clustering contextualized token embeddings into human-interpretable concept groups. We propose a hybri…

Interpretable Clustering with Adaptive Heterogeneous Causal Structure Learning in Mixed Observational Data
Wenrui Li, Qinghao Zhang, Xiaowo Wang
https://arxiv.org/abs/2509.04415

Interpretable Clustering with Adaptive Heterogeneous Causal Structure Learning in Mixed Observational Data
Understanding causal heterogeneity is essential for scientific discovery in domains such as biology and medicine. However, existing methods lack causal awareness, with insufficient modeling of heterogeneity, confounding, and observational constraints, leading to poor interpretability and difficulty distinguishing true causal heterogeneity from spurious associations. We propose an unsupervised framework, HCL (Interpretable Causal Mechanism-Aware Clustering with Adaptive Heterogeneous Causal Stru…

Architecturally Constrained Solutions to Ill-Conditioned Problems in QUBIC
Leonora Kardum
https://arxiv.org/abs/2510.00090 https://arxiv.org/pdf/2510.00090…

Architecturally Constrained Solutions to Ill-Conditioned Problems in QUBIC
This article introduces a new physics-guided Machine Learning framework, with which we solve the generally non-invertible, ill-conditioned problems through an analytical approach and constrain the solution to the approximate inverse with the architecture of Neural Networks. By informing the networks of the underlying physical processes, the method optimizes data usage and enables interpretability of the model while simultaneously allowing estimation of detector properties and the propagation of…

Interpretable Scalar-on-Image Linear Regression Models via the Generalized Dantzig Selector
Sijia Liao, Xiaoxiao Sun, Ning Hao, Hao Helen Zhang
https://arxiv.org/abs/2508.20278 …

Interpretable Scalar-on-Image Linear Regression Models via the Generalized Dantzig Selector
The scalar-on-image regression model examines the association between a scalar response and a bivariate function (e.g., images) through the estimation of a bivariate coefficient function. Existing approaches often impose smoothness constraints to control the bias-variance trade-off, and thus prevent overfitting. However, such assumptions can hinder interpretability, especially when only certain regions of an image influence changes in the response. In such a scenario, interpretability can be be…

A Neuro-Fuzzy System for Interpretable Long-Term Stock Market Forecasting
Miha O\v{z}bot, Igor \v{S}krjanc, Vitomir \v{S}truc
https://arxiv.org/abs/2510.00960 https://

A Neuro-Fuzzy System for Interpretable Long-Term Stock Market Forecasting
In the complex landscape of multivariate time series forecasting, achieving both accuracy and interpretability remains a significant challenge. This paper introduces the Fuzzy Transformer (Fuzzformer), a novel recurrent neural network architecture combined with multi-head self-attention and fuzzy inference systems to analyze multivariate stock market data and conduct long-term time series forecasting. The method leverages LSTM networks and temporal attention to condense multivariate data into i…

V-SEAM: Visual Semantic Editing and Attention Modulating for Causal Interpretability of Vision-Language Models
Qidong Wang, Junjie Hu, Ming Jiang
https://arxiv.org/abs/2509.14837

V-SEAM: Visual Semantic Editing and Attention Modulating for Causal Interpretability of Vision-Language Models
Recent advances in causal interpretability have extended from language models to vision-language models (VLMs), seeking to reveal their internal mechanisms through input interventions. While textual interventions often target semantics, visual interventions typically rely on coarse pixel-level perturbations, limiting semantic insights on multimodal integration. In this study, we introduce V-SEAM, a novel framework that combines Visual Semantic Editing and Attention Modulating for causal interpr…

AIM: Amending Inherent Interpretability via Self-Supervised Masking
Eyad Alshami, Shashank Agnihotri, Bernt Schiele, Margret Keuper
https://arxiv.org/abs/2508.11502 https://

AIM: Amending Inherent Interpretability via Self-Supervised Masking
It has been observed that deep neural networks (DNNs) often use both genuine as well as spurious features. In this work, we propose "Amending Inherent Interpretability via Self-Supervised Masking" (AIM), a simple yet interestingly effective method that promotes the network's utilization of genuine features over spurious alternatives without requiring additional annotations. In particular, AIM uses features at multiple encoding stages to guide a self-supervised, sample-specific feature-masking p…

DPsurv: Dual-Prototype Evidential Fusion for Uncertainty-Aware and Interpretable Whole-Slide Image Survival Prediction
Yucheng Xing, Ling Huang, Jingying Ma, Ruping Hong, Jiangdong Qiu, Pei Liu, Kai He, Huazhu Fu, Mengling Feng
https://arxiv.org/abs/2510.00053

DPsurv: Dual-Prototype Evidential Fusion for Uncertainty-Aware and Interpretable Whole-Slide Image Survival Prediction
Pathology whole-slide images (WSIs) are widely used for cancer survival analysis because of their comprehensive histopathological information at both cellular and tissue levels, enabling quantitative, large-scale, and prognostically rich tumor feature analysis. However, most existing methods in WSI survival analysis struggle with limited interpretability and often overlook predictive uncertainty in heterogeneous slide images. In this paper, we propose DPsurv, a dual-prototype whole-slide image …

Bayesian Neural Networks for Functional ANOVA model
Seokhun Park, Choeun Kim, Jihu Lee, Yunseop Shin, Insung Kong, Yongdai Kim
https://arxiv.org/abs/2510.00545 https://

Bayesian Neural Networks for Functional ANOVA model
With the increasing demand for interpretability in machine learning, functional ANOVA decomposition has gained renewed attention as a principled tool for breaking down high-dimensional function into low-dimensional components that reveal the contributions of different variable groups. Recently, Tensor Product Neural Network (TPNN) has been developed and applied as basis functions in the functional ANOVA model, referred to as ANOVA-TPNN. A disadvantage of ANOVA-TPNN, however, is that the compone…

Typed Chain-of-Thought: A Curry-Howard Framework for Verifying LLM Reasoning
Elija Perrier
https://arxiv.org/abs/2510.01069 https://arxiv.org/pdf/2510.0106…

Typed Chain-of-Thought: A Curry-Howard Framework for Verifying LLM Reasoning
While Chain-of-Thought (CoT) prompting enhances the reasoning capabilities of large language models, the faithfulness of the generated rationales remains an open problem for model interpretability. We propose a novel theoretical lens for this problem grounded in the Curry-Howard correspondence, which posits a direct relationship between formal proofs and computer programs. Under this paradigm, a faithful reasoning trace is analogous to a well-typed program, where each intermediate step correspo…

EvolveSignal: A Large Language Model Powered Coding Agent for Discovering Traffic Signal Control Algorithms
Leizhen Wang, Peibo Duan, Hao Wang, Yue Wang, Jian Xu, Nan Zheng, Zhenliang Ma
https://arxiv.org/abs/2509.03335

EvolveSignal: A Large Language Model Powered Coding Agent for Discovering Traffic Signal Control Algorithms
In traffic engineering, the fixed-time traffic signal control remains widely used for its low cost, stability, and interpretability. However, its design depends on hand-crafted formulas (e.g., Webster) and manual re-timing by engineers to adapt to demand changes, which is labor-intensive and often yields suboptimal results under heterogeneous or congested conditions. This paper introduces the EvolveSignal, a large language models (LLMs) powered coding agent to automatically discover new traffic…

Constrained Co-evolutionary Metamorphic Differential Testing for Autonomous Systems with an Interpretability Approach
Hossein Yousefizadeh, Shenghui Gu, Lionel C. Briand, Ali Nasr
https://arxiv.org/abs/2509.16478

Constrained Co-evolutionary Metamorphic Differential Testing for Autonomous Systems with an Interpretability Approach
Autonomous systems, such as autonomous driving systems, evolve rapidly through frequent updates, risking unintended behavioral degradations. Effective system-level testing is challenging due to the vast scenario space, the absence of reliable test oracles, and the need for practically applicable and interpretable test cases. We present CoCoMagic, a novel automated test case generation method that combines metamorphic testing, differential testing, and advanced search-based techniques to identif…

Learning Agile Gate Traversal via Analytical Optimal Policy Gradient
Tianchen Sun, Bingheng Wang, Longbin Tang, Yichao Gao, Lin Zhao
https://arxiv.org/abs/2508.21592 https://

Learning Agile Gate Traversal via Analytical Optimal Policy Gradient
Traversing narrow gates presents a significant challenge and has become a standard benchmark for evaluating agile and precise quadrotor flight. Traditional modularized autonomous flight stacks require extensive design and parameter tuning, while end-to-end reinforcement learning (RL) methods often suffer from low sample efficiency and limited interpretability. In this work, we present a novel hybrid framework that adaptively fine-tunes model predictive control (MPC) parameters online using outp…

Initialization Schemes for Kolmogorov-Arnold Networks: An Empirical Study
Spyros Rigas, Dhruv Verma, Georgios Alexandridis, Yixuan Wang
https://arxiv.org/abs/2509.03417 https://…

Initialization Schemes for Kolmogorov-Arnold Networks: An Empirical Study
Kolmogorov-Arnold Networks (KANs) are a recently introduced neural architecture that replace fixed nonlinearities with trainable activation functions, offering enhanced flexibility and interpretability. While KANs have been applied successfully across scientific and machine learning tasks, their initialization strategies remain largely unexplored. In this work, we study initialization schemes for spline-based KANs, proposing two theory-driven approaches inspired by LeCun and Glorot, as well as …

Unfolding Framework with Complex-Valued Deformable Attention for High-Quality Computer-Generated Hologram Generation
Haomiao Zhang, Zhangyuan Li, Yanling Piao, Zhi Li, Xiaodong Wang, Miao Cao, Xiongfei Su, Qiang Song, Xin Yuan
https://arxiv.org/abs/2508.21657

Unfolding Framework with Complex-Valued Deformable Attention for High-Quality Computer-Generated Hologram Generation
Computer-generated holography (CGH) has gained wide attention with deep learning-based algorithms. However, due to its nonlinear and ill-posed nature, challenges remain in achieving accurate and stable reconstruction. Specifically, ($i$) the widely used end-to-end networks treat the reconstruction model as a black box, ignoring underlying physical relationships, which reduces interpretability and flexibility. ($ii$) CNN-based CGH algorithms have limited receptive fields, hindering their ability…

MUSE-Explainer: Counterfactual Explanations for Symbolic Music Graph Classification Models
Baptiste Hilaire, Emmanouil Karystinaios, Gerhard Widmer
https://arxiv.org/abs/2509.26521

MUSE-Explainer: Counterfactual Explanations for Symbolic Music Graph Classification Models
Interpretability is essential for deploying deep learning models in symbolic music analysis, yet most research emphasizes model performance over explanation. To address this, we introduce MUSE-Explainer, a new method that helps reveal how music Graph Neural Network models make decisions by providing clear, human-friendly explanations. Our approach generates counterfactual explanations by making small, meaningful changes to musical score graphs that alter a model's prediction while ensuring the …

Latent Thinking Optimization: Your Latent Reasoning Language Model Secretly Encodes Reward Signals in its Latent Thoughts
Hanwen Du, Yuxin Dong, Xia Ning
https://arxiv.org/abs/2509.26314

Latent Thinking Optimization: Your Latent Reasoning Language Model Secretly Encodes Reward Signals in its Latent Thoughts
Large Language Models (LLMs) excel at problem solving by generating chain of thoughts in natural language, but such verbal thinking is computationally costly and prone to overthinking. Recent work instead proposes a latent thinking architecture Huggin-3.5B, which represents intermediate reasoning steps as sequence of latent representations. However, latent thoughts lack interpretability and are difficult to supervise, raising concerns about the correctness and reliability of its latent thinking…

A Neuro-Fuzzy System for Interpretable Long-Term Stock Market Forecasting
Miha O\v{z}bot, Igor \v{S}krjanc, Vitomir \v{S}truc
https://arxiv.org/abs/2510.00960 https://

A Neuro-Fuzzy System for Interpretable Long-Term Stock Market Forecasting
In the complex landscape of multivariate time series forecasting, achieving both accuracy and interpretability remains a significant challenge. This paper introduces the Fuzzy Transformer (Fuzzformer), a novel recurrent neural network architecture combined with multi-head self-attention and fuzzy inference systems to analyze multivariate stock market data and conduct long-term time series forecasting. The method leverages LSTM networks and temporal attention to condense multivariate data into i…

LINKER: Learning Interactions Between Functional Groups and Residues With Chemical Knowledge-Enhanced Reasoning and Explainability
Phuc Pham, Viet Thanh Duy Nguyen, Truong-Son Hy
https://arxiv.org/abs/2509.03425

LINKER: Learning Interactions Between Functional Groups and Residues With Chemical Knowledge-Enhanced Reasoning and Explainability
Accurate identification of interactions between protein residues and ligand functional groups is essential to understand molecular recognition and guide rational drug design. Existing deep learning approaches for protein-ligand interpretability often rely on 3D structural input or use distance-based contact labels, limiting both their applicability and biological relevance. We introduce LINKER, the first sequence-based model to predict residue-functional group interactions in terms of biologica…

Interpret, prune and distill Donut : towards lightweight VLMs for VQA on document
Adnan Ben Mansour, Ayoub Karine, David Naccache
https://arxiv.org/abs/2509.26235 https://

Interpret, prune and distill Donut : towards lightweight VLMs for VQA on document
Recent advances in Visually-rich Document Understanding rely on large Vision-Language Models like Donut, which perform document-level Visual Question Answering without Optical Character Recognition. Despite their effectiveness, these models are too costly for real-time or resource-constrained applications. We investigate model compression through knowledge distillation, training compact student models from a larger teacher. We leverage mechanistic interpretability to drive student architecture …

A more interpretable regression model for count data with excess of zeros
Gustavo H. A. Pereira, Jeremias Le\~ao, Manoel Santos-Neto, Jianwen Cai
https://arxiv.org/abs/2509.24916

A more interpretable regression model for count data with excess of zeros
Count data are common in medical research. When these data have more zeros than expected by the most used count distributions, it is common to employ a zero-inflated regression model. However, the interpretability of these models is much lower than the most used count regression models. In this work, we introduce a more interpretable regression model for count data with excess of zeros based on a reparameterization of the zero-inflated Poisson distribution. We discuss inferential and diagnostic…

Typed Chain-of-Thought: A Curry-Howard Framework for Verifying LLM Reasoning
Elija Perrier
https://arxiv.org/abs/2510.01069 https://arxiv.org/pdf/2510.0106…

Typed Chain-of-Thought: A Curry-Howard Framework for Verifying LLM Reasoning
While Chain-of-Thought (CoT) prompting enhances the reasoning capabilities of large language models, the faithfulness of the generated rationales remains an open problem for model interpretability. We propose a novel theoretical lens for this problem grounded in the Curry-Howard correspondence, which posits a direct relationship between formal proofs and computer programs. Under this paradigm, a faithful reasoning trace is analogous to a well-typed program, where each intermediate step correspo…

Hyperdimensional Probe: Decoding LLM Representations via Vector Symbolic Architectures
Marco Bronzini, Carlo Nicolini, Bruno Lepri, Jacopo Staiano, Andrea Passerini
https://arxiv.org/abs/2509.25045

Hyperdimensional Probe: Decoding LLM Representations via Vector Symbolic Architectures
Despite their capabilities, Large Language Models (LLMs) remain opaque with limited understanding of their internal representations. Current interpretability methods, such as direct logit attribution (DLA) and sparse autoencoders (SAEs), provide restricted insight due to limitations such as the model's output vocabulary or unclear feature names. This work introduces Hyperdimensional Probe, a novel paradigm for decoding information from the LLM vector space. It combines ideas from symbolic repre…

Medical priority fusion: achieving dual optimization of sensitivity and interpretability in nipt anomaly detection
Xiuqi Ge, Zhibo Yao, Yaosong Du
https://arxiv.org/abs/2509.17924

Medical priority fusion: achieving dual optimization of sensitivity and interpretability in nipt anomaly detection
Clinical machine learning faces a critical dilemma in high-stakes medical applications: algorithms achieving optimal diagnostic performance typically sacrifice the interpretability essential for physician decision-making, while interpretable methods compromise sensitivity in complex scenarios. This paradox becomes particularly acute in non-invasive prenatal testing (NIPT), where missed chromosomal abnormalities carry profound clinical consequences yet regulatory frameworks mandate explainable A…

Improving Large Language Models Function Calling and Interpretability via Guided-Structured Templates
Hy Dang, Tianyi Liu, Zhuofeng Wu, Jingfeng Yang, Haoming Jiang, Tao Yang, Pei Chen, Zhengyang Wang, Helen Wang, Huasheng Li, Bing Yin, Meng Jiang
https://arxiv.org/abs/2509.18076

Improving Large Language Models Function Calling and Interpretability via Guided-Structured Templates
Large language models (LLMs) have demonstrated strong reasoning and tool-use capabilities, yet they often fail in real-world tool-interactions due to incorrect parameterization, poor tool selection, or misinterpretation of user intent. These issues often stem from an incomplete understanding of user goals and inadequate comprehension of tool documentation. While Chain-of-Thought (CoT) prompting has proven effective for enhancing reasoning in general contexts, our analysis reveals that free-form…

UML-CoT: Structured Reasoning and Planning with Unified Modeling Language for Robotic Room Cleaning
Hongyu Chen, Guangrun Wang
https://arxiv.org/abs/2509.22628 https://

UML-CoT: Structured Reasoning and Planning with Unified Modeling Language for Robotic Room Cleaning
Chain-of-Thought (CoT) prompting improves reasoning in large language models (LLMs), but its reliance on unstructured text limits interpretability and executability in embodied tasks. Prior work has explored structured CoTs using scene or logic graphs, yet these remain fundamentally limited: they model only low-order relations, lack constructs like inheritance or behavioral abstraction, and provide no standardized semantics for sequential or conditional planning. We propose UML-CoT, a structure…

Behind the Scenes: Mechanistic Interpretability of LoRA-adapted Whisper for Speech Emotion Recognition
Yujian Ma, Jinqiu Sang, Ruizhe Li
https://arxiv.org/abs/2509.08454 https:/…

Behind the Scenes: Mechanistic Interpretability of LoRA-adapted Whisper for Speech Emotion Recognition
Large pre-trained speech models such as Whisper offer strong generalization but pose significant challenges for resource-efficient adaptation. Low-Rank Adaptation (LoRA) has become a popular parameter-efficient fine-tuning method, yet its underlying mechanisms in speech tasks remain poorly understood. In this work, we conduct the first systematic mechanistic interpretability study of LoRA within the Whisper encoder for speech emotion recognition (SER). Using a suite of analytical tools, includi…

Model Interpretability and Rationale Extraction by Input Mask Optimization
Marc Brinner, Sina Zarriess
https://arxiv.org/abs/2508.11388 https://arxiv.org/p…

Model Interpretability and Rationale Extraction by Input Mask Optimization
Concurrent to the rapid progress in the development of neural-network based models in areas like natural language processing and computer vision, the need for creating explanations for the predictions of these black-box models has risen steadily. We propose a new method to generate extractive explanations for predictions made by neural networks, that is based on masking parts of the input which the model does not consider to be indicative of the respective class. The masking is done using gradi…

Explaining multimodal LLMs via intra-modal token interactions
Jiawei Liang, Ruoyu Chen, Xianghao Jiao, Siyuan Liang, Shiming Liu, Qunli Zhang, Zheng Hu, Xiaochun Cao
https://arxiv.org/abs/2509.22415

Explaining multimodal LLMs via intra-modal token interactions
Multimodal Large Language Models (MLLMs) have achieved remarkable success across diverse vision-language tasks, yet their internal decision-making mechanisms remain insufficiently understood. Existing interpretability research has primarily focused on cross-modal attribution, identifying which image regions the model attends to during output generation. However, these approaches often overlook intra-modal dependencies. In the visual modality, attributing importance to isolated image patches ign…

Interpretability as Alignment: Making Internal Understanding a Design Principle
Aadit Sengupta, Pratinav Seth, Vinay Kumar Sankarapu
https://arxiv.org/abs/2509.08592 https://

Interpretability as Alignment: Making Internal Understanding a Design Principle
Large neural models are increasingly deployed in high-stakes settings, raising concerns about whether their behavior reliably aligns with human values. Interpretability provides a route to internal transparency by revealing the computations that drive outputs. We argue that interpretability especially mechanistic approaches should be treated as a design principle for alignment, not an auxiliary diagnostic tool. Post-hoc methods such as LIME or SHAP offer intuitive but correlational explanations…

Privacy Preserved Federated Learning with Attention-Based Aggregation for Biometric Recognition
Kassahun Azezew, Minyechil Alehegn, Tsega Asresa, Bitew Mekuria, Tizazu Bayh, Ayenew Kassie, Amsalu Tesema, Animut Embiyale
https://arxiv.org/abs/2510.01113

Privacy Preserved Federated Learning with Attention-Based Aggregation for Biometric Recognition
Because biometric data is sensitive, centralized training poses a privacy risk, even though biometric recognition is essential for contemporary applications. Federated learning (FL), which permits decentralized training, provides a privacy-preserving substitute. Conventional FL, however, has trouble with interpretability and heterogeneous data (non-IID). In order to handle non-IID biometric data, this framework adds an attention mechanism at the central server that weights local model updates a…

Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation
Ruoyu Chen, Xiaoqing Guo, Kangwei Liu, Siyuan Liang, Shiming Liu, Qunli Zhang, Hua Zhang, Xiaochun Cao
https://arxiv.org/abs/2509.22496

Where MLLMs Attend and What They Rely On: Explaining Autoregressive Token Generation
Multimodal large language models (MLLMs) have demonstrated remarkable capabilities in aligning visual inputs with natural language outputs. Yet, the extent to which generated tokens depend on visual modalities remains poorly understood, limiting interpretability and reliability. In this work, we present EAGLE, a lightweight black-box framework for explaining autoregressive token generation in MLLMs. EAGLE attributes any selected tokens to compact perceptual regions while quantifying the relativ…

REMA: A Unified Reasoning Manifold Framework for Interpreting Large Language Model
Bo Li, Guanzhi Deng, Ronghao Chen, Junrong Yue, Shuo Zhang, Qinghua Zhao, Linqi Song, Lijie Wen
https://arxiv.org/abs/2509.22518

REMA: A Unified Reasoning Manifold Framework for Interpreting Large Language Model
Understanding how Large Language Models (LLMs) perform complex reasoning and their failure mechanisms is a challenge in interpretability research. To provide a measurable geometric analysis perspective, we define the concept of the Reasoning Manifold, a latent low-dimensional geometric structure formed by the internal representations corresponding to all correctly reasoned generations. This structure can be conceptualized as the embodiment of the effective thinking paths that the model has lear…

Analysing Moral Bias in Finetuned LLMs through Mechanistic Interpretability
Bianca Raimondi, Daniela Dalbagno, Maurizio Gabbrielli
https://arxiv.org/abs/2510.12229 https://

Analysing Moral Bias in Finetuned LLMs through Mechanistic Interpretability
Large language models (LLMs) have been shown to internalize human-like biases during finetuning, yet the mechanisms by which these biases manifest remain unclear. In this work, we investigated whether the well-known Knobe effect, a moral bias in intentionality judgements, emerges in finetuned LLMs and whether it can be traced back to specific components of the model. We conducted a Layer-Patching analysis across 3 open-weights LLMs and demonstrated that the bias is not only learned during finet…

CLAIRE: A Dual Encoder Network with RIFT Loss and Phi-3 Small Language Model Based Interpretability for Cross-Modality Synthetic Aperture Radar and Optical Land Cover Segmentation
Debopom Sutradhar, Arefin Ittesafun Abian, Mohaimenul Azam Khan Raiaan, Reem E. Mohamed, Sheikh Izzal Azid, Sami Azam
https://arxiv.org/abs/2509.11952

CLAIRE: A Dual Encoder Network with RIFT Loss and Phi-3 Small Language Model Based Interpretability for Cross-Modality Synthetic Aperture Radar and Optical Land Cover Segmentation
Accurate land cover classification from satellite imagery is crucial in environmental monitoring and sustainable resource management. However, it remains challenging due to the complexity of natural landscapes, the visual similarity between classes, and the significant class imbalance in the available datasets. To address these issues, we propose a dual encoder architecture that independently extracts modality-specific features from optical and Synthetic Aperture Radar (SAR) imagery, which are …

Interpretable by AI Mother Tongue: Native Symbolic Reasoning in Neural Models
Hung Ming Liu
https://arxiv.org/abs/2508.18988 https://arxiv.org/pdf/2508.189…

Interpretable by AI Mother Tongue: Native Symbolic Reasoning in Neural Models
We present a framework where neural models develop an AI Mother Tongue, a native symbolic language that simultaneously supports intuitive reasoning, compositional symbol chains, and inherent interpretability. Unlike post-hoc explanation methods, our approach embeds reasoning directly into the model's representations: symbols capture meaningful semantic patterns, chains trace decision paths, and gated induction mechanisms guide selective focus, yielding transparent yet flexible reasoning. We int…

Towards Understanding the Shape of Representations in Protein Language Models
Kosio Beshkov, Anders Malthe-S{\o}renssen
https://arxiv.org/abs/2509.24895 https://

Towards Understanding the Shape of Representations in Protein Language Models
While protein language models (PLMs) are one of the most promising avenues of research for future de novo protein design, the way in which they transform sequences to hidden representations, as well as the information encoded in such representations is yet to be fully understood. Several works have attempted to propose interpretability tools for PLMs, but they have focused on understanding how individual sequences are transformed by such models. Therefore, the way in which PLMs transform the wh…

Tracking World States with Language Models: State-Based Evaluation Using Chess
Romain Harang, Jason Naradowsky, Yaswitha Gujju, Yusuke Miyao
https://arxiv.org/abs/2508.19851 htt…

Tracking World States with Language Models: State-Based Evaluation Using Chess
Large Language Models (LLMs) exhibit emergent capabilities in structured domains, suggesting they may implicitly internalize high-fidelity representations of world models. While probing techniques have shown promising signs of this in scientific and game-based settings, they rely on model-specific internal activations, which limit interpretability and generalizability. In this work, we propose a model-agnostic, state-based evaluation framework using chess as a benchmark to assess whether LLMs p…

RLBFF: Binary Flexible Feedback to bridge between Human Feedback & Verifiable Rewards
Zhilin Wang, Jiaqi Zeng, Olivier Delalleau, Ellie Evans, Daniel Egert, Hoo-Chang Shin, Felipe Soares, Yi Dong, Oleksii Kuchaiev
https://arxiv.org/abs/2509.21319

RLBFF: Binary Flexible Feedback to bridge between Human Feedback & Verifiable Rewards
Reinforcement Learning with Human Feedback (RLHF) and Reinforcement Learning with Verifiable Rewards (RLVR) are the main RL paradigms used in LLM post-training, each offering distinct advantages. However, RLHF struggles with interpretability and reward hacking because it relies on human judgments that usually lack explicit criteria, whereas RLVR is limited in scope by its focus on correctness-based verifiers. We propose Reinforcement Learning with Binary Flexible Feedback (RLBFF), which combine…

(Sometimes) Less is More: Mitigating the Complexity of Rule-based Representation for Interpretable Classification
Luca Bergamin, Roberto Confalonieri, Fabio Aiolli
https://arxiv.org/abs/2509.22384

(Sometimes) Less is More: Mitigating the Complexity of Rule-based Representation for Interpretable Classification
Deep neural networks are widely used in practical applications of AI, however, their inner structure and complexity made them generally not easily interpretable. Model transparency and interpretability are key requirements for multiple scenarios where high performance is not enough to adopt the proposed solution. In this work, a differentiable approximation of $L_0$ regularization is adapted into a logic-based neural network, the Multi-layer Logical Perceptron (MLLP), to study its efficacy in r…

MOSS-ChatV: Reinforcement Learning with Process Reasoning Reward for Video Temporal Reasoning
Sicheng Tao, Jungang Li, Yibo Yan, Junyan Zhang, Yubo Gao, Hanqian Li, ShuHang Xun, Yuxuan Fan, Hong Chen, Jianxiang He, Xuming Hu
https://arxiv.org/abs/2509.21113

MOSS-ChatV: Reinforcement Learning with Process Reasoning Reward for Video Temporal Reasoning
Video reasoning has emerged as a critical capability for multimodal large language models (MLLMs), requiring models to move beyond static perception toward coherent understanding of temporal dynamics in complex scenes. Yet existing MLLMs often exhibit process inconsistency, where intermediate reasoning drifts from video dynamics even when the final answer is correct, undermining interpretability and robustness. To address this issue, we introduce MOSS-ChatV, a reinforcement learning framework w…

Enhancing Credit Risk Prediction: A Meta-Learning Framework Integrating Baseline Models, LASSO, and ECOC for Superior Accuracy
Haibo Wang, Lutfu S. Sua, Jun Huang, Figen Balo, Burak Dolar
https://arxiv.org/abs/2509.22381

Enhancing Credit Risk Prediction: A Meta-Learning Framework Integrating Baseline Models, LASSO, and ECOC for Superior Accuracy
Effective credit risk management is fundamental to financial decision-making, necessitating robust models for default probability prediction and financial entity classification. Traditional machine learning approaches face significant challenges when confronted with high-dimensional data, limited interpretability, rare event detection, and multi-class imbalance problems in risk assessment. This research proposes a comprehensive meta-learning framework that synthesizes multiple complementary mod…

Interpretable Decision-Making for End-to-End Autonomous Driving
Mona Mirzaie, Bodo Rosenhahn
https://arxiv.org/abs/2508.18898 https://arxiv.org/pdf/2508.18…

Interpretable Decision-Making for End-to-End Autonomous Driving
Trustworthy AI is mandatory for the broad deployment of autonomous vehicles. Although end-to-end approaches derive control commands directly from raw data, interpreting these decisions remains challenging, especially in complex urban scenarios. This is mainly attributed to very deep neural networks with non-linear decision boundaries, making it challenging to grasp the logic behind AI-driven decisions. This paper presents a method to enhance interpretability while optimizing control commands in…

Think as a Doctor: An Interpretable AI Approach for ICU Mortality Prediction
Qingwen Li, Xiaohang Zhao, Xiao Han, Hailiang Huang, Lanjuan Liu
https://arxiv.org/abs/2510.11745 ht…

Think as a Doctor: An Interpretable AI Approach for ICU Mortality Prediction
Intensive Care Unit (ICU) mortality prediction, which estimates a patient's mortality status at discharge using EHRs collected early in an ICU admission, is vital in critical care. For this task, predictive accuracy alone is insufficient; interpretability is equally essential for building clinical trust and meeting regulatory standards, a topic that has attracted significant attention in information system research. Accordingly, an ideal solution should enable intrinsic interpretability and ali…