arXiv subjects

Alberto Gasparin

Publications and source records attributed to Alberto Gasparin.

7 recordsLinked to original sources

arXiv2026

Post-hoc Alignment of LLM-judges to Human Judgment Distribution

The LLM-as-a-judge (LLMaJ) framework offers a cost-effective and reproducible solution for automatic evaluation. However, current evaluation practices typically compare LLMaJ judgments against aggregated ground-truth labels, overlooking the valuable information contained in Human Label Variation (HLV). Inspired by an increasing line of work that proposes to leverage HLV, we systematically study LLMaJ performance on predicting both a single, aggregated ground truth hard-label and unaggregated soft-labels that represent Human Judgment Distributions (HJD). Our results across five diverse datasets reveal that while LLMs achieve near human-level performance at hard-label prediction on most tasks, they exhibit poor performance when predicting soft-labels. To address this limitation, we propose NAPHA (eNtropy-Aware Post-Hoc Alignment), a simple yet effective lightweight post-hoc alignment method that matches the LLM distribution to the HJD by first assigning an instance to a discrete entropy class and then routing it to specialized, trained alignment models. We find that NAPHA consistently improves soft-labels prediction across base LLM models and datasets, with particularly strong gains on high-entropy instances where capturing diverse human perspectives is most critical. We also show via oracle experiments that improving entropy class prediction can substantially enhance NAPHA's practical effectiveness.

cs.CL↗

arXiv2026

Does task decomposition improve automatic NLG evaluation?

The LLM-as-a-judge (LLMaJ) framework has emerged as a promising solution for cheap, reproducible, reference-free Natural Language Generation (NLG) evaluation. Prior work seeks to improve LLMaJ by decomposing evaluation tasks into simpler sub-tasks. In this work, we systematically compare LLMaJ methods with and without decomposition on multiple NLG datasets. We find no evidence that LLMaJ with task decomposition leads to performance gains over a fair baseline that does not use decomposition. Instead, we find that previously reported performance gains in decomposition-based LLMaJ stem from using human labels as training data, and not task decomposition itself. Also, we find that, when human labels are available, LLMaJ without using task decomposition can perform comparably to human annotators.

cs.CL↗

arXiv2023

Fortuna: A Library for Uncertainty Quantification in Deep Learning

We present Fortuna, an open-source library for uncertainty quantification in deep learning. Fortuna supports a range of calibration techniques, such as conformal prediction that can be applied to any trained neural network to generate reliable uncertainty estimates, and scalable Bayesian inference methods that can be applied to Flax-based deep neural networks trained from scratch for improved uncertainty quantification and accuracy. By providing a coherent framework for advanced uncertainty quantification methods, Fortuna simplifies the process of benchmarking and helps practitioners build robust AI systems.

cs.LG↗

arXiv2022

Uncertainty Calibration in Bayesian Neural Networks via Distance-Aware Priors

As we move away from the data, the predictive uncertainty should increase, since a great variety of explanations are consistent with the little available information. We introduce Distance-Aware Prior (DAP) calibration, a method to correct overconfidence of Bayesian deep learning models outside of the training domain. We define DAPs as prior distributions over the model parameters that depend on the inputs through a measure of their distance from the training set. DAP calibration is agnostic to the posterior inference method, and it can be performed as a post-processing step. We demonstrate its effectiveness against several baselines in a variety of classification and regression problems, including benchmarks designed to test the quality of predictive distributions away from the data.

stat.ML↗

arXiv2021

Ranking Micro-Influencers: a Novel Multi-Task Learning and Interpretable Framework

With the rise in use of social media to promote branded products, the demand for effective influencer marketing has increased. Brands are looking for improved ways to identify valuable influencers among a vast catalogue; this is even more challenging with "micro-influencers", which are more affordable than mainstream ones but difficult to discover. In this paper, we propose a novel multi-task learning framework to improve the state of the art in micro-influencer ranking based on multimedia content. Moreover, since the visual congruence between a brand and influencer has been shown to be good measure of compatibility, we provide an effective visual method for interpreting our models' decisions, which can also be used to inform brands' media strategies. We compare with the current state-of-the-art on a recently constructed public dataset and we show significant improvement both in terms of accuracy and model complexity. The techniques for ranking and interpretation presented in this work can be generalised to arbitrary multimedia ranking tasks that have datasets with a similar structure.

cs.LG↗

arXiv2019

Deep Learning for Time Series Forecasting: The Electric Load Case

Management and efficient operations in critical infrastructure such as Smart Grids take huge advantage of accurate power load forecasting which, due to its nonlinear nature, remains a challenging task. Recently, deep learning has emerged in the machine learning field achieving impressive performance in a vast range of tasks, from image classification to machine translation. Applications of deep learning models to the electric load forecasting problem are gaining interest among researchers as well as the industry, but a comprehensive and sound comparison among different architectures is not yet available in the literature. This work aims at filling the gap by reviewing and experimentally evaluating on two real-world datasets the most recent trends in electric load forecasting, by contrasting deep learning architectures on short term forecast (one day ahead prediction). Specifically, we focus on feedforward and recurrent neural networks, sequence to sequence models and temporal convolutional neural networks along with architectural variants, which are known in the signal processing community but are novel to the load forecasting one.

cs.LG↗

arXiv2018

Deriving item features relevance from collaborative domain knowledge

An Item based recommender system works by computing a similarity between items, which can exploit past user interactions (collaborative filtering) or item features (content based filtering). Collaborative algorithms have been proven to achieve better recommendation quality then content based algorithms in a variety of scenarios, being more effective in modeling user behaviour. However, they can not be applied when items have no interactions at all, i.e. cold start items. Content based algorithms, which are applicable to cold start items, often require a lot of feature engineering in order to generate useful recommendations. This issue is specifically relevant as the content descriptors become large and heterogeneous. The focus of this paper is on how to use a collaborative models domain-specific knowledge to build a wrapper feature weighting method which embeds collaborative knowledge in a content based algorithm. We present a comparative study for different state of the art algorithms and present a more general model. This machine learning approach to feature weighting shows promising results and high flexibility.

cs.IR↗