arXiv subjects

Elena Baralis

Publications and source records attributed to Elena Baralis.

25 records · Page 2Linked to original sources

arXiv2023

Explaining Speech Classification Models via Word-Level Audio Segments and Paralinguistic Features

Recent advances in eXplainable AI (XAI) have provided new insights into how models for vision, language, and tabular data operate. However, few approaches exist for understanding speech models. Existing work focuses on a few spoken language understanding (SLU) tasks, and explanations are difficult to interpret for most users. We introduce a new approach to explain speech classification models. We generate easy-to-interpret explanations via input perturbation on two information levels. 1) Word-level explanations reveal how each word-related audio segment impacts the outcome. 2) Paralinguistic features (e.g., prosody and background noise) answer the counterfactual: ``What would the model prediction be if we edited the audio signal in this way?'' We validate our approach by explaining two state-of-the-art SLU models on two speech classification tasks in English and Italian. Our findings demonstrate that the explanations are faithful to the model's inner workings and plausible to humans. Our method and findings pave the way for future research on interpreting speech models.

cs.CL↗

arXiv2023

ITALIC: An Italian Intent Classification Dataset

Recent large-scale Spoken Language Understanding datasets focus predominantly on English and do not account for language-specific phenomena such as particular phonemes or words in different lects. We introduce ITALIC, the first large-scale speech dataset designed for intent classification in Italian. The dataset comprises 16,521 crowdsourced audio samples recorded by 70 speakers from various Italian regions and annotated with intent labels and additional metadata. We explore the versatility of ITALIC by evaluating current state-of-the-art speech and text models. Results on intent classification suggest that increasing scale and running language adaptation yield better speech models, monolingual text models outscore multilingual ones, and that speech recognition on ITALIC is more challenging than on existing Italian benchmarks. We release both the dataset and the annotation scheme to streamline the development of new Italian SLU models and language-specific datasets.

cs.CL↗

arXiv2022

Entropy-based Attention Regularization Frees Unintended Bias Mitigation from Lists

Natural Language Processing (NLP) models risk overfitting to specific terms in the training data, thereby reducing their performance, fairness, and generalizability. E.g., neural hate speech detection models are strongly influenced by identity terms like gay, or women, resulting in false positives, severe unintended bias, and lower performance. Most mitigation techniques use lists of identity terms or samples from the target domain during training. However, this approach requires a-priori knowledge and introduces further bias if important terms are neglected. Instead, we propose a knowledge-free Entropy-based Attention Regularization (EAR) to discourage overfitting to training-specific terms. An additional objective function penalizes tokens with low self-attention entropy. We fine-tune BERT via EAR: the resulting model matches or exceeds state-of-the-art performance for hate speech classification and bias metrics on three benchmark corpora in English and Italian. EAR also reveals overfitting terms, i.e., terms most likely to induce bias, to help identify their effect on the model, task, and predictions.

cs.CL↗

arXiv2021

Identifying Biased Subgroups in Ranking and Classification

When analyzing the behavior of machine learning algorithms, it is important to identify specific data subgroups for which the considered algorithm shows different performance with respect to the entire dataset. The intervention of domain experts is normally required to identify relevant attributes that define these subgroups. We introduce the notion of divergence to measure this performance difference and we exploit it in the context of (i) classification models and (ii) ranking applications to automatically detect data subgroups showing a significant deviation in their behavior. Furthermore, we quantify the contribution of all attributes in the data subgroup to the divergent behavior by means of Shapley values, thus allowing the identification of the most impacting attributes.

cs.LG↗

arXiv2019

Automating concept-drift detection by self-evaluating predictive model degradation

A key aspect of automating predictive machine learning entails the capability of properly triggering the update of the trained model. To this aim, suitable automatic solutions to self-assess the prediction quality and the data distribution drift between the original training set and the new data have to be devised. In this paper, we propose a novel methodology to automatically detect prediction-quality degradation of machine learning models due to class-based concept drift, i.e., when new data contains samples that do not fit the set of class labels known by the currently-trained predictive model. Experiments on synthetic and real-world public datasets show the effectiveness of the proposed methodology in automatically detecting and describing concept drift caused by changes in the class-label data distributions.

cs.LG↗

arXiv2018

Scaling associative classification for very large datasets

Supervised learning algorithms are nowadays successfully scaling up to datasets that are very large in volume, leveraging the potential of in-memory cluster-computing Big Data frameworks. Still, massive datasets with a number of large-domain categorical features are a difficult challenge for any classifier. Most off-the-shelf solutions cannot cope with this problem. In this work we introduce DAC, a Distributed Associative Classifier. DAC exploits ensemble learning to distribute the training of an associative classifier among parallel workers and improve the final quality of the model. Furthermore, it adopts several novel techniques to reach high scalability without sacrificing quality, among which a preventive pruning of classification rules in the extraction phase based on Gini impurity. We ran experiments on Apache Spark, on a real large-scale dataset with more than 4 billion records and 800 million distinct categories. The results showed that DAC improves on a state-of-the-art solution in both prediction quality and execution time. Since the generated model is human-readable, it can not only classify new records, but also allow understanding both the logic behind the prediction and the properties of the model, becoming a useful aid for decision makers.

cs.LG↗

arXiv2015

YouLighter: An Unsupervised Methodology to Unveil YouTube CDN Changes

YouTube relies on a massively distributed Content Delivery Network (CDN) to stream the billions of videos in its catalogue. Unfortunately, very little information about the design of such CDN is available. This, combined with the pervasiveness of YouTube, poses a big challenge for Internet Service Providers (ISPs), which are compelled to optimize end-users' Quality of Experience (QoE) while having no control on the CDN decisions. This paper presents YouLighter, an unsupervised technique to identify changes in the YouTube CDN. YouLighter leverages only passive measurements to cluster co-located identical caches into edge-nodes. This automatically unveils the structure of YouTube's CDN. Further, we propose a new metric, called Constellation Distance, that compares the clustering obtained from two different time snapshots, to pinpoint sudden changes. While several approaches allow comparison between the clustering results from the same dataset, no technique allows to measure the similarity of clusters from different datasets. Hence, we develop a novel methodology, based on the Constellation Distance, to solve this problem. By running YouLighter over 10-month long traces obtained from two ISPs in different countries, we pinpoint both sudden changes in edge-node allocation, and small alterations to the cache allocation policies which actually impair the QoE that the end-users perceive.

cs.NI↗