arXiv subjects

Edoardo Serra

Publications and source records attributed to Edoardo Serra.

8 recordsLinked to original sources

arXiv2026

Testing LLM Arithmetic Reasoning Generalization with Automatic Numeric-Remapping Attacks

Large language models achieve strong performance on arithmetic reasoning benchmarks, and one common response to arithmetic brittleness is to delegate computation to code. Yet models are still often used in settings where they must reason directly from natural language, and trustworthy models should solve small-number arithmetic word problems without external tools. Prior work shows that LLMs are sensitive to numerical variation: a model may solve an original problem but fail on structurally similar variants requiring the same reasoning procedure with different numbers. We ask whether this fragility persists under a stricter setting involving small, schema-preserving numeric changes that retain the original reasoning program and avoid large-number stress tests. We introduce an automatic algorithm for generating numeric-remapping attacks on arithmetic word problems. Unlike template-based perturbation methods requiring manual schemas or constraints, our approach derives problem-specific symbolic representations, generates constrained numeric remappings, recomputes gold answers, and realizes transformed questions through deterministic edits guided by LLM-generated edit plans. Stage-wise validation and a high-confidence audit retain reliable attacks, making the pipeline scalable with limited human intervention. We evaluate DeepSeek-R1 (70B), Gemma4 (31B), and GPT-OSS (120B) on GSM8K, MAWPS, and MultiArith. On GSM8K, completed runs show conditional accuracy drops of 12.16 to 25.82 percentage points. MAWPS and MultiArith are far more stable, with most attacked accuracies near or above 98%. These results show that numeric-remapping robustness depends strongly on dataset structure: GSM8K remains sensitive even when reasoning programs are preserved and answers are recomputed, while shorter, more regular datasets are more robust.

cs.CR↗

arXiv2026

Scaling Novel Graph Generation via Lightweight Structure-Guided Autoregressive Models

Generating realistic and diverse graphs is a key problem in machine learning, with applications in molecular discovery, circuit design, cybersecurity, and beyond. However, current graph generative models remain limited by scalability and novelty. Diffusion-based methods often require costly full-adjacency operations and long denoising chains, while many autoregressive and hybrid models have at least quadratic complexity. In addition, these models often imitate training graphs rather than generalize beyond them. We propose a lightweight autoregressive framework to address these issues. It uses a structure-guided topological ordering to serialize graphs into regular edge sequences, enabling near log-linear generation, and a two-phase training strategy that combines exploration-oriented augmentation with iterative refinement to reduce overfitting and promote controlled novelty. Experiments on molecular and non-molecular benchmarks show that our approach improves novelty while preserving high validity and uniqueness. The framework also supports both LSTM and Mamba-style causal sequence backbones, with large-memory accelerators enabling longer graph-sequence experiments beyond typical GPU limits.

cs.LG↗

arXiv2024

AGS-GNN: Attribute-guided Sampling for Graph Neural Networks

We propose AGS-GNN, a novel attribute-guided sampling algorithm for Graph Neural Networks (GNNs) that exploits node features and connectivity structure of a graph while simultaneously adapting for both homophily and heterophily in graphs. (In homophilic graphs vertices of the same class are more likely to be connected, and vertices of different classes tend to be linked in heterophilic graphs.) While GNNs have been successfully applied to homophilic graphs, their application to heterophilic graphs remains challenging. The best-performing GNNs for heterophilic graphs do not fit the sampling paradigm, suffer high computational costs, and are not inductive. We employ samplers based on feature-similarity and feature-diversity to select subsets of neighbors for a node, and adaptively capture information from homophilic and heterophilic neighborhoods using dual channels. Currently, AGS-GNN is the only algorithm that we know of that explicitly controls homophily in the sampled subgraph through similar and diverse neighborhood samples. For diverse neighborhood sampling, we employ submodularity, which was not used in this context prior to our work. The sampling distribution is pre-computed and highly parallel, achieving the desired scalability. Using an extensive dataset consisting of 35 small ($\le$ 100K nodes) and large (>100K nodes) homophilic and heterophilic graphs, we demonstrate the superiority of AGS-GNN compare to the current approaches in the literature. AGS-GNN achieves comparable test accuracy to the best-performing heterophilic GNNs, even outperforming methods using the entire graph for node classification. AGS-GNN also converges faster compared to methods that sample neighborhoods randomly, and can be incorporated into existing GNN models that employ node or graph sampling.

cs.LG↗

arXiv2024

GenFighter: A Generative and Evolutive Textual Attack Removal

Adversarial attacks pose significant challenges to deep neural networks (DNNs) such as Transformer models in natural language processing (NLP). This paper introduces a novel defense strategy, called GenFighter, which enhances adversarial robustness by learning and reasoning on the training classification distribution. GenFighter identifies potentially malicious instances deviating from the distribution, transforms them into semantically equivalent instances aligned with the training data, and employs ensemble techniques for a unified and robust response. By conducting extensive experiments, we show that GenFighter outperforms state-of-the-art defenses in accuracy under attack and attack success rate metrics. Additionally, it requires a high number of queries per attack, making the attack more challenging in real scenarios. The ablation study shows that our approach integrates transfer learning, a generative/evolutive procedure, and an ensemble method, providing an effective defense against NLP adversarial attacks.

cs.LG↗

arXiv2021

Inferential SIR-GN: Scalable Graph Representation Learning

Graph representation learning methods generate numerical vector representations for the nodes in a network, thereby enabling their use in standard machine learning models. These methods aim to preserve relational information, such that nodes that are similar in the graph are found close to one another in the representation space. Similarity can be based largely on one of two notions: connectivity or structural role. In tasks where node structural role is important, connectivity based methods show poor performance. Recent work has begun to focus on scalability of learning methods to massive graphs of millions to billions of nodes and edges. Many unsupervised node representation learning algorithms are incapable of scaling to large graphs, and are unable to generate node representations for unseen nodes. In this work, we propose Inferential SIR-GN, a model which is pre-trained on random graphs, then computes node representations rapidly, including for very large networks. We demonstrate that the model is able to capture node's structural role information, and show excellent performance at node and graph classification tasks, on unseen networks. Additionally, we observe the scalability of Inferential SIR-GN is comparable to the fastest current approaches for massive graphs.

cs.LG↗

arXiv2021

V2W-BERT: A Framework for Effective Hierarchical Multiclass Classification of Software Vulnerabilities

Weaknesses in computer systems such as faults, bugs and errors in the architecture, design or implementation of software provide vulnerabilities that can be exploited by attackers to compromise the security of a system. Common Weakness Enumerations (CWE) are a hierarchically designed dictionary of software weaknesses that provide a means to understand software flaws, potential impact of their exploitation, and means to mitigate these flaws. Common Vulnerabilities and Exposures (CVE) are brief low-level descriptions that uniquely identify vulnerabilities in a specific product or protocol. Classifying or mapping of CVEs to CWEs provides a means to understand the impact and mitigate the vulnerabilities. Since manual mapping of CVEs is not a viable option, automated approaches are desirable but challenging. We present a novel Transformer-based learning framework (V2W-BERT) in this paper. By using ideas from natural language processing, link prediction and transfer learning, our method outperforms previous approaches not only for CWE instances with abundant data to train, but also rare CWE classes with little or no data to train. Our approach also shows significant improvements in using historical data to predict links for future instances of CVEs, and therefore, provides a viable approach for practical applications. Using data from MITRE and National Vulnerability Database, we achieve up to 97% prediction accuracy for randomly partitioned data and up to 94% prediction accuracy in temporally partitioned data. We believe that our work will influence the design of better methods and training models, as well as applications to solve increasingly harder problems in cybersecurity.

cs.LG↗

arXiv2016

ChoiceGAPs: Competitive Diffusion as a Massive Multi-Player Game in Social Networks

We consider the problem of modeling competitive diffusion in real world social networks via the notion of ChoiceGAPs which combine choice logic programs due to Sacca` and Zaniolo and Generalized Annotated Programs due to Kifer and Subrahmanian. We assume that each vertex in a social network is a player in a multi-player game (with a huge number of players) - the choice part of the ChoiceGAPs describe utilities of players for acting in various ways based on utilities of their neighbors in those and other situations. We define multi-player Nash equilibrium for such programs - but because they require some conditions that are hard to satisfy in the real world, we introduce a new model-theoretic concept of strong equilibrium. We show that stable equilibria can capture all Nash equilibria. We prove a host of complexity (intractability) results for checking existence of strong equilibria (as well as related counting complexity results), together with algorithms to find them. We then identify a class of ChoiceGAPs for which stable equilibria can be polynomially computed. We develop algorithms for computing these equilibria under various restrictions. We come up with the important concept of an estimation query which can compute quantities w.r.t. a given strong equilibrium, and approximate ranges of values (answers) across the space of strong equilibria. Even though we show that computing range answers to estimation queries exactly is intractable, we are able to identify classes of estimation queries that can be answered in polynomial time. We report on experiments we conducted with a real-world FaceBook data set surrounding the 2013 Italian election showing that our algorithms have good predictive accuracy with an Area Under a ROC Curve that, on average, is over 0.76.

cs.LO↗

arXiv2013

Multi-Sorted Inverse Frequent Itemsets Mining

The development of novel platforms and techniques for emerging "Big Data" applications requires the availability of real-life datasets for data-driven experiments, which are however out of reach for academic research in most cases as they are typically proprietary. A possible solution is to use synthesized datasets that reflect patterns of real ones in order to ensure high quality experimental findings. A first step in this direction is to use inverse mining techniques such as inverse frequent itemset mining (IFM) that consists of generating a transactional database satisfying given support constraints on the itemsets in an input set, that are typically the frequent ones. This paper introduces an extension of IFM, called many-sorted IFM, where the schemes for the datasets to be generated are those typical of Big Tables as required in emerging big data applications, e.g., social network analytics.

cs.DB↗