Normalized AOPC: Fixing Misleading Faithfulness Metrics for Feature Attribution Explainability

Deep neural network predictions are notoriously difficult to interpret. Feature attribution methods aim to explain these predictions by identifying the contribution of each input feature. Faithfulness, often evaluated using the area over the perturbation curve (AOPC), reflects feature attributions' accuracy in describing the internal mechanisms of deep neural networks. However, many studies rely on AOPC to compare faithfulness across different models, which we show can lead to false conclusions about models' faithfulness. Specifically, we find that AOPC is sensitive to variations in the model, resulting in unreliable cross-model comparisons. Moreover, AOPC scores are difficult to interpret in isolation without knowing the model-specific lower and upper limits. To address these issues, we propose a normalization approach, Normalized AOPC (NAOPC), enabling consistent cross-model evaluations and more meaningful interpretation of individual scores. Our experiments demonstrate that this normalization can radically change AOPC results, questioning the conclusions of earlier studies and offering a more robust framework for assessing feature attribution faithfulness.

From Internal Conflict to Contextual Adaptation of Language Models

Knowledge-intensive language understanding tasks require Language Models (LMs) to integrate relevant context, mitigating their inherent weaknesses, such as incomplete or outdated knowledge. Nevertheless, studies indicate that LMs often ignore the provided context as it can conflict with the pre-existing LM's memory learned during pre-training. Moreover, conflicting knowledge can already be present in the LM's parameters, termed intra-memory conflict. Existing works have studied the two types of knowledge conflicts only in isolation. We conjecture that the (degree of) intra-memory conflicts can in turn affect LM's handling of context-memory conflicts. To study this, we introduce the DYNAMICQA dataset, which includes facts with a temporal dynamic nature where a fact can change with a varying time frequency and disputable dynamic facts, which can change depending on the viewpoint. DYNAMICQA is the first to include real-world knowledge conflicts and provide context to study the link between the different types of knowledge conflicts. With the proposed dataset, we assess the use of uncertainty for measuring the intra-memory conflict and introduce a novel Coherent Persuasion (CP) score to evaluate the context's ability to sway LM's semantic output. Our extensive experiments reveal that static facts, which are unlikely to change, are more easily updated with additional context, relative to temporal and disputable facts.

An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Records

Electronic healthcare records are vital for patient safety as they document conditions, plans, and procedures in both free text and medical codes. Language models have significantly enhanced the processing of such records, streamlining workflows and reducing manual data entry, thereby saving healthcare providers significant resources. However, the black-box nature of these models often leaves healthcare professionals hesitant to trust them. State-of-the-art explainability methods increase model transparency but rely on human-annotated evidence spans, which are costly. In this study, we propose an approach to produce plausible and faithful explanations without needing such annotations. We demonstrate on the automated medical coding task that adversarial robustness training improves explanation plausibility and introduce AttInGrad, a new explanation method superior to previous ones. By combining both contributions in a fully unsupervised setup, we produce explanations of comparable quality, or better, to that of a supervised approach. We release our code and model weights.

Can We Trust Recommender System Fairness Evaluation? The Role of Fairness and Relevance

Relevance and fairness are two major objectives of recommender systems (RSs). Recent work proposes measures of RS fairness that are either independent from relevance (fairness-only) or conditioned on relevance (joint measures). While fairness-only measures have been studied extensively, we look into whether joint measures can be trusted. We collect all joint evaluation measures of RS relevance and fairness, and ask: How much do they agree with each other? To what extent do they agree with relevance/fairness measures? How sensitive are they to changes in rank position, or to increasingly fair and relevant recommendations? We empirically study for the first time the behaviour of these measures across 4 real-world datasets and 4 recommenders. We find that most of these measures: i) correlate weakly with one another and even contradict each other at times; ii) are less sensitive to rank position changes than relevance- and fairness-only measures, meaning that they are less granular than traditional RS measures; and iii) tend to compress scores at the low end of their range, meaning that they are not very expressive. We counter the above limitations with a set of guidelines on the appropriate usage of such measures, i.e., they should be used with caution due to their tendency to contradict each other and of having a very small empirical range.

Dataset and Models for Item Recommendation Using Multi-Modal User Interactions

While recommender systems with multi-modal item representations (image, audio, and text), have been widely explored, learning recommendations from multi-modal user interactions (e.g., clicks and speech) remains an open problem. We study the case of multi-modal user interactions in a setting where users engage with a service provider through multiple channels (website and call center). In such cases, incomplete modalities naturally occur, since not all users interact through all the available channels. To address these challenges, we publish a real-world dataset that allows progress in this under-researched area. We further present and benchmark various methods for leveraging multi-modal user interactions for item recommendations, and propose a novel approach that specifically deals with missing modalities by mapping user interactions to a common feature space. Our analysis reveals important interactions between the different modalities and that a frequently occurring modality can enhance learning from a less frequent one.

Recommending Target Actions Outside Sessions in the Data-poor Insurance Domain

Providing personalized recommendations for insurance products is particularly challenging due to the intrinsic and distinctive features of the insurance domain. First, unlike more traditional domains like retail, movie etc., a large amount of user feedback is not available and the item catalog is smaller. Second, due to the higher complexity of products, the majority of users still prefer to complete their purchases over the phone instead of online. We present different recommender models to address such data scarcity in the insurance domain. We use recurrent neural networks with 3 different types of loss functions and architectures (cross-entropy, censored Weibull, attention). Our models cope with data scarcity by learning from multiple sessions and different types of user actions. Moreover, differently from previous session-based models, our models learn to predict a target action that does not happen within the session. Our models outperform state-of-the-art baselines on a real-world insurance dataset, with ca. 44K users, 16 items, 54K purchases and 117K sessions. Moreover, combining our models with demographic data boosts the performance. Analysis shows that considering multiple sessions and several types of actions are both beneficial for the models, and that our models are not unfair with respect to age, gender and income.

Evaluation Measures of Individual Item Fairness for Recommender Systems: A Critical Study

Fairness is an emerging and challenging topic in recommender systems. In recent years, various ways of evaluating and therefore improving fairness have emerged. In this study, we examine existing evaluation measures of fairness in recommender systems. Specifically, we focus solely on exposure-based fairness measures of individual items that aim to quantify the disparity in how individual items are recommended to users, separate from item relevance to users. We gather all such measures and we critically analyse their theoretical properties. We identify a series of limitations in each of them, which collectively may render the affected measures hard or impossible to interpret, to compute, or to use for comparing recommendations. We resolve these limitations by redefining or correcting the affected measures, or we argue why certain limitations cannot be resolved. We further perform a comprehensive empirical analysis of both the original and our corrected versions of these fairness measures, using real-world and synthetic datasets. Our analysis provides novel insights into the relationship between measures based on different fairness concepts, and different levels of measure sensitivity and strictness. We conclude with practical suggestions of which fairness measures should be used and when. Our code is publicly available. To our knowledge, this is the first critical comparison of individual item fairness measures in recommender systems.

Automated Medical Coding on MIMIC-III and MIMIC-IV: A Critical Review and Replicability Study

Medical coding is the task of assigning medical codes to clinical free-text documentation. Healthcare professionals manually assign such codes to track patient diagnoses and treatments. Automated medical coding can considerably alleviate this administrative burden. In this paper, we reproduce, compare, and analyze state-of-the-art automated medical coding machine learning models. We show that several models underperform due to weak configurations, poorly sampled train-test splits, and insufficient evaluation. In previous work, the macro F1 score has been calculated sub-optimally, and our correction doubles it. We contribute a revised model comparison using stratified sampling and identical experimental setups, including hyperparameters and decision boundary tuning. We analyze prediction errors to validate and falsify assumptions of previous works. The analysis confirms that all models struggle with rare codes, while long documents only have a negligible impact. Finally, we present the first comprehensive results on the newly released MIMIC-IV dataset using the reproduced models. We release our code, model parameters, and new MIMIC-III and MIMIC-IV training and evaluation pipelines to accommodate fair future comparisons.

Graph-based Recommendation for Sparse and Heterogeneous User Interactions

Recommender system research has oftentimes focused on approaches that operate on large-scale datasets containing millions of user interactions. However, many small businesses struggle to apply state-of-the-art models due to their very limited availability of data. We propose a graph-based recommender model which utilizes heterogeneous interactions between users and content of different types and is able to operate well on small-scale datasets. A genetic algorithm is used to find optimal weights that represent the strength of the relationship between users and content. Experiments on two real-world datasets (which we make available to the research community) show promising results (up to 7% improvement), in comparison with other state-of-the-art methods for low-data environments. These improvements are statistically significant and consistent across different data samples.

Principled Multi-Aspect Evaluation Measures of Rankings

Information Retrieval evaluation has traditionally focused on defining principled ways of assessing the relevance of a ranked list of documents with respect to a query. Several methods extend this type of evaluation beyond relevance, making it possible to evaluate different aspects of a document ranking (e.g., relevance, usefulness, or credibility) using a single measure (multi-aspect evaluation). However, these methods either are (i) tailor-made for specific aspects and do not extend to other types or numbers of aspects, or (ii) have theoretical anomalies, e.g. assign maximum score to a ranking where all documents are labelled with the lowest grade with respect to all aspects (e.g., not relevant, not credible, etc.). We present a theoretically principled multi-aspect evaluation method that can be used for any number, and any type, of aspects. A thorough empirical evaluation using up to 5 aspects and a total of 425 runs officially submitted to 10 TREC tracks shows that our method is more discriminative than the state-of-the-art and overcomes theoretical limitations of the state-of-the-art.