arXiv subjects

Ozlem Ozmen Garibay

Publications and source records attributed to Ozlem Ozmen Garibay.

18 recordsLinked to original sources

arXiv2026

Smarter Saboteurs, Better Fixers: Scaling & Security in Linear Multi-Agent Workflows

As LLM-based multi-agent systems (MAS) are deployed in the wild, the resilience of their collaboration structures against adversarial compromise becomes a critical safety concern. Attackers may leverage prompt-injection or jailbreaking to sabotage individual agents within MAS workflows, but the interaction between model scaling and system-level resilience remains poorly understood. This paper investigates how model scale affects the security of linear multi-agent workflows. Our experiments across scales of two open-weight model families on the HumanEval benchmark reveal a compliance-correction symmetry: larger models are far more likely to faithfully execute malicious instructions, with the control-to-malicious performance drop reaching 53.7pp at 27B in uncorrected pipelines. However, appending a lightweight terminal Fixer stage collapses this to 0.6pp and restores statistical parity with control-level performance, demonstrating that strictly linear collaboration structures can be viable and resilient to adversaries at this scale, and suggesting that the brittleness previously attributed to linear topology may stem from a lack of correction.

cs.MA↗

arXiv2026

ConRetroBert: EMA Stabilized Dual Encoders for Template-Based Single-Step Retrosynthesis

Template based single step retrosynthesis predicts reactants by selecting and applying an explicit reaction template, making each prediction traceable to a chemical transformation rule. This is useful for synthesis planning, but template based methods are often viewed as less competitive than template free models because template prediction is commonly formulated as global classification over a long tailed rule library. We argue that this weakness is not inherent to templates, but to the learning formulation. We present ConRetroBert, a dual encoder framework that reframes template based retrosynthesis as dense product template retrieval followed by candidate set listwise ranking. Stage 1 uses contrastive pretraining to learn a shared embedding space between products and reaction templates. Stage 2 refines template ranking over mined hard negative candidate sets with a multi positive listwise objective. To enable template side adaptation without destabilizing hard negative mining, ConRetroBert uses a slow moving exponential moving average template encoder for retrieval bank construction while updating the live template encoder through the ranking loss. On the local USPTO-50k benchmark, Stage 2 candidate set ranking improves top-1 reaction accuracy from 50.5% to 61.3%, while EMA stabilized template adaptation further improves it to 62.4%. Fine tuning from a leakage controlled USPTO-Full checkpoint reaches 75.4% top-1 accuracy on USPTO-50k. We also show that retrieval based template prediction is strong in the long tail of rare templates, and that many correct reactant predictions arise from alternative explicit templates rather than only the recorded positive label. Code and data are available at https://github.com/JahidBasher/ConRetroBert.

cs.LG↗

arXiv2025

FairContrast: Enhancing Fairness through Contrastive learning and Customized Augmenting Methods on Tabular Data

As AI systems become more embedded in everyday life, the development of fair and unbiased models becomes more critical. Considering the social impact of AI systems is not merely a technical challenge but a moral imperative. As evidenced in numerous research studies, learning fair and robust representations has proven to be a powerful approach to effectively debiasing algorithms and improving fairness while maintaining essential information for prediction tasks. Representation learning frameworks, particularly those that utilize self-supervised and contrastive learning, have demonstrated superior robustness and generalizability across various domains. Despite the growing interest in applying these approaches to tabular data, the issue of fairness in these learned representations remains underexplored. In this study, we introduce a contrastive learning framework specifically designed to address bias and learn fair representations in tabular datasets. By strategically selecting positive pair samples and employing supervised and self-supervised contrastive learning, we significantly reduce bias compared to existing state-of-the-art contrastive learning models for tabular data. Our results demonstrate the efficacy of our approach in mitigating bias with minimum trade-off in accuracy and leveraging the learned fair representations in various downstream tasks.

cs.LG↗

arXiv2025

Equi-mRNA: Protein Translation Equivariant Encoding for mRNA Language Models

The growing importance of mRNA therapeutics and synthetic biology highlights the need for models that capture the latent structure of synonymous codon (different triplets encoding the same amino acid) usage, which subtly modulates translation efficiency and gene expression. While recent efforts incorporate codon-level inductive biases through auxiliary objectives, they often fall short of explicitly modeling the structured relationships that arise from the genetic code's inherent symmetries. We introduce Equi-mRNA, the first codon-level equivariant mRNA language model that explicitly encodes synonymous codon symmetries as cyclic subgroups of 2D Special Orthogonal matrix (SO(2)). By combining group-theoretic priors with an auxiliary equivariance loss and symmetry-aware pooling, Equi-mRNA learns biologically grounded representations that outperform vanilla baselines across multiple axes. On downstream property-prediction tasks including expression, stability, and riboswitch switching Equi-mRNA delivers up to approximately 10% improvements in accuracy. In sequence generation, it produces mRNA constructs that are up to approximately 4x more realistic under Frechet BioDistance metrics and approximately 28% better preserve functional properties compared to vanilla baseline. Interpretability analyses further reveal that learned codon-rotation distributions recapitulate known GC-content biases and tRNA abundance patterns, offering novel insights into codon usage. Equi-mRNA establishes a new biologically principled paradigm for mRNA modeling, with significant implications for the design of next-generation therapeutics.

q-bio.QM↗

arXiv2025

LLM-Mixer: Multiscale Mixing in LLMs for Time Series Forecasting

Time series forecasting remains a challenging task, particularly in the context of complex multiscale temporal patterns. This study presents LLM-Mixer, a framework that improves forecasting accuracy through the combination of multiscale time-series decomposition with pre-trained LLMs (Large Language Models). LLM-Mixer captures both short-term fluctuations and long-term trends by decomposing the data into multiple temporal resolutions and processing them with a frozen LLM, guided by a textual prompt specifically designed for time-series data. Extensive experiments conducted on multivariate and univariate datasets demonstrate that LLM-Mixer achieves competitive performance, outperforming recent state-of-the-art models across various forecasting horizons. This work highlights the potential of combining multiscale analysis and LLMs for effective and scalable time-series forecasting.

cs.LG↗

arXiv2025

Does Self-Attention Need Separate Weights in Transformers?

The success of self-attention lies in its ability to capture long-range dependencies and enhance context understanding, but it is limited by its computational complexity and challenges in handling sequential data with inherent directionality. This work introduces a shared weight self-attention-based BERT model that only learns one weight matrix for (Key, Value, and Query) representations instead of three individual matrices for each of them. Our shared weight attention reduces the training parameter size by more than half and training time by around one-tenth. Furthermore, we demonstrate higher prediction accuracy on small tasks of GLUE over the BERT baseline and in particular a generalization power on noisy and out-of-domain data. Experimental results indicate that our shared self-attention method achieves a parameter size reduction of 66.53% in the attention block. In the GLUE dataset, the shared weight self-attention-based BERT model demonstrates accuracy improvements of 0.38%, 5.81%, and 1.06% over the standard, symmetric, and pairwise attention-based BERT models, respectively. The model and source code are available at Anonymous.

cs.CL↗

arXiv2025

PEFT A2Z: Parameter-Efficient Fine-Tuning Survey for Large Language and Vision Models

Large models such as Large Language Models (LLMs) and Vision Language Models (VLMs) have transformed artificial intelligence, powering applications in natural language processing, computer vision, and multimodal learning. However, fully fine-tuning these models remains expensive, requiring extensive computational resources, memory, and task-specific data. Parameter-Efficient Fine-Tuning (PEFT) has emerged as a promising solution that allows adapting large models to downstream tasks by updating only a small portion of parameters. This survey presents a comprehensive overview of PEFT techniques, focusing on their motivations, design principles, and effectiveness. We begin by analyzing the resource and accessibility challenges posed by traditional fine-tuning and highlight key issues, such as overfitting, catastrophic forgetting, and parameter inefficiency. We then introduce a structured taxonomy of PEFT methods -- grouped into additive, selective, reparameterized, hybrid, and unified frameworks -- and systematically compare their mechanisms and trade-offs. Beyond taxonomy, we explore the impact of PEFT across diverse domains, including language, vision, and generative modeling, showing how these techniques offer strong performance with lower resource costs. We also discuss important open challenges in scalability, interpretability, and robustness, and suggest future directions such as federated learning, domain adaptation, and theoretical grounding. Our goal is to provide a unified understanding of PEFT and its growing role in enabling practical, efficient, and sustainable use of large models.

cs.CL↗

arXiv2025

BnTTS: Few-Shot Speaker Adaptation in Low-Resource Setting

This paper introduces BnTTS (Bangla Text-To-Speech), the first framework for Bangla speaker adaptation-based TTS, designed to bridge the gap in Bangla speech synthesis using minimal training data. Building upon the XTTS architecture, our approach integrates Bangla into a multilingual TTS pipeline, with modifications to account for the phonetic and linguistic characteristics of the language. We pre-train BnTTS on 3.85k hours of Bangla speech dataset with corresponding text labels and evaluate performance in both zero-shot and few-shot settings on our proposed test dataset. Empirical evaluations in few-shot settings show that BnTTS significantly improves the naturalness, intelligibility, and speaker fidelity of synthesized Bangla speech. Compared to state-of-the-art Bangla TTS systems, BnTTS exhibits superior performance in Subjective Mean Opinion Score (SMOS), Naturalness, and Clarity metrics.

cs.CL↗

arXiv2025

BoKDiff: Best-of-K Diffusion Alignment for Target-Specific 3D Molecule Generation

Structure-based drug design (SBDD) leverages the 3D structure of biomolecular targets to guide the creation of new therapeutic agents. Recent advances in generative models, including diffusion models and geometric deep learning, have demonstrated promise in optimizing ligand generation. However, the scarcity of high-quality protein-ligand complex data and the inherent challenges in aligning generated ligands with target proteins limit the effectiveness of these methods. We propose BoKDiff, a novel framework that enhances ligand generation by combining multi-objective optimization and Best-of-K alignment methodologies. Built upon the DecompDiff model, BoKDiff generates diverse candidates and ranks them using a weighted evaluation of molecular properties such as QED, SA, and docking scores. To address alignment challenges, we introduce a method that relocates the center of mass of generated ligands to their docking poses, enabling accurate sub-component extraction. Additionally, we integrate a Best-of-N (BoN) sampling approach, which selects the optimal ligand from multiple generated candidates without requiring fine-tuning. BoN achieves exceptional results, with QED values exceeding 0.6, SA scores above 0.75, and a success rate surpassing 35%, demonstrating its efficiency and practicality. BoKDiff achieves state-of-the-art results on the CrossDocked2020 dataset, including a -8.58 average Vina docking score and a 26% success rate in molecule generation. This study is the first to apply Best-of-K alignment and Best-of-N sampling to SBDD, highlighting their potential to bridge generative modeling with practical drug discovery requirements. The code is provided at https://github.com/khodabandeh-ali/BoKDiff.git.

q-bio.BM↗

arXiv2024

Fair Bilevel Neural Network (FairBiNN): On Balancing fairness and accuracy via Stackelberg Equilibrium

The persistent challenge of bias in machine learning models necessitates robust solutions to ensure parity and equal treatment across diverse groups, particularly in classification tasks. Current methods for mitigating bias often result in information loss and an inadequate balance between accuracy and fairness. To address this, we propose a novel methodology grounded in bilevel optimization principles. Our deep learning-based approach concurrently optimizes for both accuracy and fairness objectives, and under certain assumptions, achieving proven Pareto optimal solutions while mitigating bias in the trained model. Theoretical analysis indicates that the upper bound on the loss incurred by this method is less than or equal to the loss of the Lagrangian approach, which involves adding a regularization term to the loss function. We demonstrate the efficacy of our model primarily on tabular datasets such as UCI Adult and Heritage Health. When benchmarked against state-of-the-art fairness methods, our model exhibits superior performance, advancing fairness-aware machine learning solutions and bridging the accuracy-fairness gap. The implementation of FairBiNN is available on https://github.com/yazdanimehdi/FairBiNN.

cs.LG↗

arXiv2024

Parameter-Efficient Fine-Tuning of Large Language Models using Semantic Knowledge Tuning

Large Language Models (LLMs) are gaining significant popularity in recent years for specialized tasks using prompts due to their low computational cost. Standard methods like prefix tuning utilize special, modifiable tokens that lack semantic meaning and require extensive training for best performance, often falling short. In this context, we propose a novel method called Semantic Knowledge Tuning (SK-Tuning) for prompt and prefix tuning that employs meaningful words instead of random tokens. This method involves using a fixed LLM to understand and process the semantic content of the prompt through zero-shot capabilities. Following this, it integrates the processed prompt with the input text to improve the model's performance on particular tasks. Our experimental results show that SK-Tuning exhibits faster training times, fewer parameters, and superior performance on tasks such as text classification and understanding compared to other tuning methods. This approach offers a promising method for optimizing the efficiency and effectiveness of LLMs in processing language tasks.

cs.CL↗

arXiv2024

Analyzing X's Web of Influence: Dissecting News Sharing Dynamics through Credibility and Popularity with Transfer Entropy and Multiplex Network Measures

The dissemination of news articles on social media platforms significantly impacts the public's perception of global issues, with the nature of these articles varying in credibility and popularity. The challenge of measuring this influence and identifying key propagators is formidable. Traditional graph-based metrics such as different centrality measures and node degree methods offer some insights into information flow but prove insufficient for identifying hidden influencers in large-scale social media networks such as X (previously known as Twitter). This study adopts and enhances a non-parametric framework based on Transfer Entropy to elucidate the influence relationships among X users. It further categorizes the distribution of influence exerted by these actors through the innovative use of multiplex network measures within a social media context, aiming to pinpoint influential actors during significant world events. The methodology was applied to three distinct events, and the findings revealed that actors in different events leveraged different types of news articles and influenced distinct sets of actors based on the news category. Notably, we found that actors disseminating trustworthy news articles to influence others occasionally resort to untrustworthy sources. However, the converse scenario, wherein actors predominantly using untrustworthy news types switch to trustworthy sources for influence, is less prevalent. This asymmetry suggests a discernible pattern in the strategic use of news articles for influence across social media networks, highlighting the nuanced roles of trustworthiness and popularity in the spread of information and influence.

cs.SI↗

arXiv2024

Agent-Based Modeling of C. Difficile Spread in Hospitals: Assessing Contribution of High-Touch vs. Low-Touch Surfaces and Inoculations' Containment Impact

Health issues and pandemics remain paramount concerns in the contemporary era. Clostridioides Difficile Infection (CDI) stands out as a critical healthcare-associated infection with global implications. Effectively understanding the mechanisms of infection dissemination within healthcare units and hospitals is imperative to implement targeted containment measures. In this study, we address the limitations of prior research by Sulyok et al., where they delineated two distinct categories of surfaces as high-touch and low-touch fomites, and subsequently evaluated the viral spread contribution of each surface utilizing mathematical modeling and Ordinary Differential Equations (ODE). Acknowledging the indispensable role of spatial features and heterogeneity in the modeling of hospital and healthcare settings, we employ agent-based modeling to capture new insights. By incorporating spatial considerations and heterogeneous patients, we explore the impact of high-touch and low-touch surfaces on contamination transmission between patients. Furthermore, the study encompasses a comprehensive assessment of various cleaning protocols, with differing intervals and detergent cleaning efficacies, in order to identify the most optimal cleaning strategy and the most important factor amidst the array of alternatives. Our results indicate that, among various factors, the frequency of cleaning intervals is the most critical element for controlling the spread of CDI in a hospital environment.

cs.MA↗

arXiv2024

Controlling the Misinformation Diffusion in Social Media by the Effect of Different Classes of Agents

The rapid and widespread dissemination of misinformation through social networks is a growing concern in today's digital age. This study focused on modeling fake news diffusion, discovering the spreading dynamics, and designing control strategies. A common approach for modeling the misinformation dynamics is SIR-based models. Our approach is an extension of a model called 'SBFC' which is a SIR-based model. This model has three states, Susceptible, Believer, and Fact-Checker. The dynamics and transition between states are based on neighbors' beliefs, hoax credibility, spreading rate, probability of verifying the news, and probability of forgetting the current state. Our contribution is to push this model to real social networks by considering different classes of agents with their characteristics. We proposed two main strategies for confronting misinformation diffusion. First, we can educate a minor class, like scholars or influencers, to improve their ability to verify the news or remember their state longer. The second strategy is adding fact-checker bots to the network to spread the facts and influence their neighbors' states. Our result shows that both of these approaches can effectively control the misinformation spread.

cs.MA↗

arXiv2023

Distraction is All You Need for Fairness

Bias in training datasets must be managed for various groups in classification tasks to ensure parity or equal treatment. With the recent growth in artificial intelligence models and their expanding role in automated decision-making, ensuring that these models are not biased is vital. There is an abundance of evidence suggesting that these models could contain or even amplify the bias present in the data on which they are trained, inherent to their objective function and learning algorithms; Many researchers direct their attention to this issue in different directions, namely, changing data to be statistically independent, adversarial training for restricting the capabilities of a particular competitor who aims to maximize parity, etc. These methods result in information loss and do not provide a suitable balance between accuracy and fairness or do not ensure limiting the biases in training. To this end, we propose a powerful strategy for training deep learning models called the Distraction module, which can be theoretically proven effective in controlling bias from affecting the classification results. This method can be utilized with different data types (e.g., Tabular, images, graphs, etc.). We demonstrate the potency of the proposed method by testing it on UCI Adult and Heritage Health datasets (tabular), POKEC-Z, POKEC-N and NBA datasets (graph), and CelebA dataset (vision). Using state-of-the-art methods proposed in the fairness literature for each dataset, we exhibit our model is superior to these proposed methods in minimizing bias and maintaining accuracy.

cs.LG↗

arXiv2023

FragXsiteDTI: Revealing Responsible Segments in Drug-Target Interaction with Transformer-Driven Interpretation

Drug-Target Interaction (DTI) prediction is vital for drug discovery, yet challenges persist in achieving model interpretability and optimizing performance. We propose a novel transformer-based model, FragXsiteDTI, that aims to address these challenges in DTI prediction. Notably, FragXsiteDTI is the first DTI model to simultaneously leverage drug molecule fragments and protein pockets. Our information-rich representations for both proteins and drugs offer a detailed perspective on their interaction. Inspired by the Perceiver IO framework, our model features a learnable latent array, initially interacting with protein binding site embeddings using cross-attention and later refined through self-attention and used as a query to the drug fragments in the drug's cross-attention transformer block. This learnable query array serves as a mediator and enables seamless information translation, preserving critical nuances in drug-protein interactions. Our computational results on three benchmarking datasets demonstrate the superior predictive power of our model over several state-of-the-art models. We also show the interpretability of our model in terms of the critical components of both target proteins and drug molecules within drug-target pairs.

cs.LG↗

arXiv2022

Through a fair looking-glass: mitigating bias in image datasets

With the recent growth in computer vision applications, the question of how fair and unbiased they are has yet to be explored. There is abundant evidence that the bias present in training data is reflected in the models, or even amplified. Many previous methods for image dataset de-biasing, including models based on augmenting datasets, are computationally expensive to implement. In this study, we present a fast and effective model to de-bias an image dataset through reconstruction and minimizing the statistical dependence between intended variables. Our architecture includes a U-net to reconstruct images, combined with a pre-trained classifier which penalizes the statistical dependence between target attribute and the protected attribute. We evaluate our proposed model on CelebA dataset, compare the results with a state-of-the-art de-biasing method, and show that the model achieves a promising fairness-accuracy combination.

cs.CV↗

arXiv2021

TabFairGAN: Fair Tabular Data Generation with Generative Adversarial Networks

With the increasing reliance on automated decision making, the issue of algorithmic fairness has gained increasing importance. In this paper, we propose a Generative Adversarial Network for tabular data generation. The model includes two phases of training. In the first phase, the model is trained to accurately generate synthetic data similar to the reference dataset. In the second phase we modify the value function to add fairness constraint, and continue training the network to generate data that is both accurate and fair. We test our results in both cases of unconstrained, and constrained fair data generation. In the unconstrained case, i.e. when the model is only trained in the first phase and is only meant to generate accurate data following the same joint probability distribution of the real data, the results show that the model beats state-of-the-art GANs proposed in the literature to produce synthetic tabular data. Also, in the constrained case in which the first phase of training is followed by the second phase, we train the network and test it on four datasets studied in the fairness literature and compare our results with another state-of-the-art pre-processing method, and present the promising results that it achieves. Comparing to other studies utilizing GANs for fair data generation, our model is comparably more stable by using only one critic, and also by avoiding major problems of original GAN model, such as mode-dropping and non-convergence, by implementing a Wasserstein GAN.

cs.LG↗