Flavio Figueiredo

Felipe Giori, Flavio Figueiredo

Pairwise Causal Discovery is the task of determining causal, anticausal, confounded or independence relationships from pairs of variables. Over the last few years, this challenging task has promoted not only the discovery of novel machine learning models aimed at solving the task, but also discussions on how learning the causal direction of variables may benefit machine learning overall. In this paper, we show that Quantitative Information Flow (QIF), a measure usually employed for measuring leakages of information from a system to an attacker, shows promising results as features for the task. In particular, experiments with real-world datasets indicate that QIF is statistically tied to the state of the art. Our initial results motivate further inquiries on how QIF relates to causality and what are its limitations.

David Perera, Victor Moura, Lais Isabelle Alves dos Santos et al.

Characterizing precisely the asymptotic generalization error of neural networks using parameters that can be estimated efficiently is a crucial problem in machine learning, which relies heavily on heuristics and practitioners' intuition to make key design choices. In order to mitigate this issue, we introduce the Representation Gap, a metric closely related to the generalization error, but admitting better-behaved asymptotic dynamics. Focusing on equivariant diffusion models and leveraging results from optimal quantization and point-process theory, we derive a precise asymptotic equivalent of the Representation Gap and show that it is governed by a single parameter, the \textit{intrinsic dimension} of the task, which is easy to interpret, efficient to estimate, and can be linked to the equivariances of common neural network architectures. We show that this asymptotic dynamic also extends to a broader range of tasks and training algorithms. Finally, we demonstrate empirically that our asymptotic law and intrinsic dimension estimation are accurate on a wide range of synthetic datasets, where these quantities are known, as well as on more realistic datasets, where we obtain results consistent with the related literature.

Guilherme H. Resende, Luiz F. Nery, Fabrício Benevenuto et al.

Language is a dynamic aspect of our culture that changes when expressed in different technologies/communities. Online social networks have enabled the diffusion and evolution of different dialects, including African American English (AAE). However, this increased usage is not without barriers. One particular barrier is how sentiment (Vader, TextBlob, and Flair) and toxicity (Google's Perspective and the open-source Detoxify) methods present biases towards utterances with AAE expressions. Consider Google's Perspective to understand bias. Here, an utterance such as ``All n*ggers deserve to die respectfully. The police murder us.'' it reaches a higher toxicity than ``African-Americans deserve to die respectfully. The police murder us.''. This score difference likely arises because the tool cannot understand the re-appropriation of the term ``n*gger''. One explanation for this bias is that AI models are trained on limited datasets, and using such a term in training data is more likely to appear in a toxic utterance. While this may be plausible, the tool will make mistakes regardless. Here, we study bias on two Web-based (YouTube and Twitter) datasets and two spoken English datasets. Our analysis shows how most models present biases towards AAE in most settings. We isolate the impact of AAE expression usage via linguistic control features from the Linguistic Inquiry and Word Count (LIWC) software, grammatical control features extracted via Part-of-Speech (PoS) tagging from Natural Language Processing (NLP) models, and the semantic of utterances by comparing sentence embeddings from recent language models. We present consistent results on how a heavy usage of AAE expressions may cause the speaker to be considered substantially more toxic, even when speaking about nearly the same subject. Our study complements similar analyses focusing on small datasets and/or one method only.

Flavio Figueiredo, José Geraldo Fernandes, Jackson Silva et al.

Copulas are powerful statistical tools for capturing dependencies across data dimensions. Applying Copulas involves estimating independent marginals, a straightforward task, followed by the much more challenging task of determining a single copulating function, $C$, that links these marginals. For bivariate data, a copula takes the form of a two-increasing function $C: (u,v)\in \mathbb{I}^2 \rightarrow \mathbb{I}$, where $\mathbb{I} = [0, 1]$. This paper proposes 2-Cats, a Neural Network (NN) model that learns two-dimensional Copulas without relying on specific Copula families (e.g., Archimedean). Furthermore, via both theoretical properties of the model and a Lagrangian training approach, we show that 2-Cats meets the desiderata of Copula properties. Moreover, inspired by the literature on Physics-Informed Neural Networks and Sobolev Training, we further extend our training strategy to learn not only the output of a Copula but also its derivatives. Our proposed method exhibits superior performance compared to the state-of-the-art across various datasets while respecting (provably for most and approximately for a single other) properties of C.

Gabriel Souza, Flavio Figueiredo, Alexei Machado et al.

In recent years, deep learning has achieved formidable results in creative computing. When it comes to music, one viable model for music generation are Transformer based models. However, while transformers models are popular for music generation, they often rely on annotated structural information. In this work, we inquire if the off-the-shelf Music Transformer models perform just as well on structural similarity metrics using only unannotated MIDI information. We show that a slight tweak to the most common representation yields small but significant improvements. We also advocate that searching for better unannotated musical representations is more cost-effective than producing large amounts of curated and annotated data.

Flavio Figueiredo, Giovanni Martinelli, Henrique Sousa et al.

Recent advances in AI music (AIM) generation services are currently transforming the music industry. Given these advances, understanding how humans perceive AIM is crucial both to educate users on identifying AIM songs, and, conversely, to improve current models. We present results from a listener-focused experiment aimed at understanding how humans perceive AIM. In a blind, Turing-like test, participants were asked to distinguish, from a pair, the AIM and human-made song. We contrast with other studies by utilizing a randomized controlled crossover trial that controls for pairwise similarity and allows for a causal interpretation. We are also the first study to employ a novel, author-uncontrolled dataset of AIM songs from real-world usage of commercial models (i.e., Suno). We establish that listeners' reliability in distinguishing AIM causally increases when pairs are similar. Lastly, we conduct a mixed-methods content analysis of listeners' free-form feedback, revealing a focus on vocal and technical cues in their judgments.

Tales Panoutsos, Rodrygo L. T. Santos, Flavio Figueiredo

In this paper, we introduce Symmetric Low-Rank Adapters, an optimized variant of LoRA with even fewer weights. This method utilizes Low-Rank Symmetric Weight Matrices to learn downstream tasks more efficiently. Traditional LoRA accumulates fine-tuning weights with the original pre-trained weights via a Singular Value Decomposition (SVD) like approach, i.e., model weights are fine-tuned via updates of the form $BA$ (where $B \in \mathbb{R}^{n\times r}$, $A \in \mathbb{R}^{r\times n}$, and $r$ is the rank of the merged weight matrix). In contrast, our approach, named SymLoRA, represents fine-tuning weights as a Spectral Decomposition, i.e., $Q \, diag(Λ)\, Q^T$, where $Q \in \mathbb{R}^{n\times r}$ and $Λ\in \mathbb{R}^r$. SymLoRA requires approximately half of the finetuning weights. Here, we show that this approach has negligible losses in downstream efficacy.

Flavio Figueiredo, Guilherme Borges, Pedro O. S. Vaz de Melo et al.

We here focus on the task of learning Granger causality matrices for multivariate point processes. In order to accomplish this task, our work is the first to explore the use of Wold processes. By doing so, we are able to develop asymptotically fast MCMC learning algorithms. With $N$ being the total number of events and $K$ the number of processes, our learning algorithm has a $O(N(\,\log(N)\,+\,\log(K)))$ cost per iteration. This is much faster than the $O(N^3\,K^2)$ or $O(K^3)$ for the state of the art. Our approach, called GrangerBusca, is validated on nine datasets. This is an advance in relation to most prior efforts which focus mostly on subsets of the Memetracker data. Regarding accuracy, GrangerBusca is three times more accurate (in Precision@10) than the state of the art for the commonly explored subsets Memetracker. Due to GrangerBusca's much lower training complexity, our approach is the only one able to train models for larger, full, sets of data.

Flavio Figueiredo, Bruno Ribeiro, Jussara Almeida et al.

Which song will Smith listen to next? Which restaurant will Alice go to tomorrow? Which product will John click next? These applications have in common the prediction of user trajectories that are in a constant state of flux over a hidden network (e.g. website links, geographic location). What users are doing now may be unrelated to what they will be doing in an hour from now. Mindful of these challenges we propose TribeFlow, a method designed to cope with the complex challenges of learning personalized predictive models of non-stationary, transient, and time-heterogeneous user trajectories. TribeFlow is a general method that can perform next product recommendation, next song recommendation, next location prediction, and general arbitrary-length user trajectory prediction without domain-specific knowledge. TribeFlow is more accurate and up to 413x faster than top competitors.

Elizeu Santos-Neto, Flavio Figueiredo, Nigini Oliveira et al.

Tagging is a popular feature that supports several collaborative tasks, including search, as tags produced by one user can help others finding relevant content. However, task performance depends on the existence of 'good' tags. A first step towards creating incentives for users to produce 'good' tags is the quantification of their value in the first place. This work fills this gap by combining qualitative and quantitative research methods. In particular, using contextual interviews, we first determine aspects that influence users' perception of tags' value for exploratory search. Next, we formalize some of the identified aspects and propose an information-theoretical method with provable properties that quantifies the two most important aspects (according to the qualitative analysis) that influence the perception of tag value: the ability of a tag to reduce the search space while retrieving relevant items to the user. The evaluation on real data shows that our method is accurate: tags that users consider more important have higher value than tags users have not expressed interest.

Flavio Figueiredo, Jussara M. Almeida, Marcos André Gonçalves et al.

We here focus on the problem of predicting the popularity trend of user generated content (UGC) as early as possible. Taking YouTube videos as case study, we propose a novel two-step learning approach that: (1) extracts popularity trends from previously uploaded objects, and (2) predicts trends for new content. Unlike previous work, our solution explicitly addresses the inherent tradeoff between prediction accuracy and remaining interest in the content after prediction, solving it on a per-object basis. Our experimental results show great improvements of our solution over alternatives, and its applicability to improve the accuracy of state-of-the-art popularity prediction methods.