论文检索

Huifang Sun, Jiaming Pei, Lukun Wang

To diminish the substantial communication costs incurred by federated learning during the training of the global model and enhance the model update efficiency across both clients and server domains, we have integrated knowledge distillation into the federated learning framework. This integration has led to the development of a novel approach termed ClientsToServerKDFL, which streamlines the distillation process by directly transferring model insights from clients to the server for computational learning without the need for extensive computations across numerous clients. This iterative process ensures model accuracy and curtails communication expenses. Experimental data analysis has validated the efficacy of this algorithm.

Pirzada Suhail, Amit Sethi

Neural networks have emerged as powerful tools across various applications, yet their decision-making process often remains opaque, leading to them being perceived as "black boxes." This opacity raises concerns about their interpretability and reliability, especially in safety-critical scenarios. Network inversion techniques offer a solution by allowing us to peek inside these black boxes, revealing the features and patterns learned by the networks behind their decision-making processes and thereby provide valuable insights into how neural networks arrive at their conclusions, making them more interpretable and trustworthy. This paper presents a simple yet effective approach to network inversion using a meticulously conditioned generator that learns the data distribution in the input space of the trained neural network, enabling the reconstruction of inputs that would most likely lead to the desired outputs. To capture the diversity in the input space for a given output, instead of simply revealing the conditioning labels to the generator, we encode the conditioning label information into vectors and intermediate matrices and further minimize the cosine similarity between features of the generated images.

Laven Srivastava, Ishaan Gakhar

Semantic segmentation of marine environments is essential for autonomous navigation of unmanned surface vessels (USVs) as well as the detection of environmental hazards such as oil spills. To tackle the challenges of accurate environmental perception, we propose a lightweight semantic segmentation network, LAqua (Laplacians for Aquatic Segmentation), which leverages Laplacian pyramids to enhance edge detection in marine imagery. Our method drastically reduces computational requirements while maintaining high accuracy in generating semantic masks for marine environments. We evaluate LAqua on two distinct datasets: one focused on detecting oil spills in port environments and another on environmental segmentation for USVs. Results show that LAqua not only performs well across varied marine settings but also achieves comparable or superior segmentation accuracy with far fewer parameters than other models. This efficiency highlights LAqua's potential for applications in real-time detection for marine environments.

Shovito Barua Soumma, Abdullah Mamun, Hassan Ghasemzadeh

FuSE-MET addresses critical challenges in deploying human activity recognition (HAR) systems in uncontrolled environments by effectively managing noisy labels, sparse data, and undefined activity vocabularies. By integrating BERT-based word embeddings with domain-specific knowledge (i.e., MET values), FuSE-MET optimizes label merging, reducing label complexity and improving classification accuracy. Our approach outperforms the state-of-the-art techniques, including ChatGPT-4, by balancing semantic meaning and physical intensity.

Kartik Singhal, Gautam Shroff

The Abstraction and Reasoning Corpus (ARC) poses a significant challenge to artificial intelligence, demanding broad generalization and few-shot learning capabilities that remain elusive for current deep learning methods, including large language models (LLMs) (Chollet 2019). While LLMs excel in program synthesis, their direct application to ARC yields limited success. To address this, we introduce ConceptSearch, a novel function-search algorithm that leverages LLMs for program generation and employs a concept-based scoring method to guide the search efficiently. Experimental results demonstrate that ConceptSearch outperforms direct GPT-4 prompting, with our novel scoring function boosting efficiency by ~30% compared to the baseline Hamming distance scoring. Code at https://github.com/kksinghal/concept-search

Dong Shu, Mengnan Du

Demonstration selection algorithms play a crucial role in optimizing Large Language Models' (LLMs) in-context learning performance. Despite numerous proposed algorithms, their comparative effectiveness remains understudied. We present a comprehensive evaluation of six state-of-the-art demonstration selection algorithms across five datasets, examining both their effectiveness and computational efficiency. Our findings reveal significant trade-offs: while some demonstration selection algorithms achieve superior accuracy, they incur substantial computational costs. We also discover that increasing demonstration examples doesn't consistently improve performance, and some sophisticated algorithms struggle to outperform random selection in certain scenarios. These insights provide valuable benchmarks for future algorithm development and practical implementation. Our code is available at https://github.com/Tizzzzy/Demonstration_Selection_Overview.

Zitong Shen, Kangzhong Wang, Youqian Zhang, Grace Ngai, Eugene Yujun Fu

Phone scams pose a significant threat to individuals and communities, causing substantial financial losses and emotional distress. Despite ongoing efforts to combat these scams, scammers continue to adapt and refine their tactics, making it imperative to explore innovative countermeasures. This research explores the potential of large language models (LLMs) to provide detection of fraudulent phone calls. By analyzing the conversational dynamics between scammers and victims, LLM-based detectors can identify potential scams as they occur, offering immediate protection to users. While such approaches demonstrate promising results, we also acknowledge the challenges of biased datasets, relatively low recall, and hallucinations that must be addressed for further advancement in this field.

Akshat Santhana Gopalan, Sowmya Ramaswamy Krishnan

This paper investigates the application of Generative Flow Networks (GFlowNets) to lead optimization in drug discovery. GFlowNets provide a novel framework for generating diverse molecular structures while optimizing for desired properties, addressing the limitations of traditional methods in exploring vast chemical spaces. We adapt GFlowNets to incrementally modify lead compounds, integrating domain-specific heuristics to guide the generation process. Our method employs the trajectory balance objective on a graph neural network (GNN), to learn a policy that samples fragments based on a multi-objective reward. The reward function ensures increase in cell permeability and similarity to the starting molecule. The results on benchmark datasets of activity cliffs demonstrate that GFlowNets can generate diverse modifications, producing optimized candidate molecules with improvement in cell permeability. This work can be extended with other pharmacokinetic properties for lead optimization in early-stage drug development, potentially accelerating the discovery of novel therapeutics.

Nilanjana Saha, Narayan Changder, Redha Taguelmimt, Samir Aknine, Animesh Dutta

On social media, it is easy to see how people are connected and find the leader, or mastermind of a network. The mastermind is responsible for the planning of the activities in the network. Hiding the mastermind is important to carry out these activities. This raises the question for the mastermind: How effectively can the mastermind hide his connections to avoid being found? We propose an efficient heuristic algorithm called HERMES (Hide Exposures by Removing Mastermind’s External Sources) to address this. Experiments on Facebook and Google networks show that HERMES hides the mastermind more effectively than the state-of-the-art, achieving time gains of 103 and 1397 seconds, respectively, and improving influence value by up to 11.11%.

Emily Jimin Roh, Joo Yong Shim, Soohyun Park, Joongheon Kim

This paper proposes a novel quantum style transfer (QST) in hybrid quantum-classical computing. QST leverages quantum computing's ability to process high-dimensional data efficiently. Our approach aims to decrease both inference time and complexity while maintaining performance, presenting a viable solution that enhances the scalability and efficiency of image generation technologies.

Caroline Rinks

Intimate Partner Violence is a global, life-threatening public health issue that can be prevented by recognizing emotionally aggressive behaviors that signal the potential for future relationship abuse. To help identify these precursory unhealthy behaviors, this study proposes a Multi-task Learning framework for training robust models capable of detecting not only physically abusive behaviors but also emotionally abusive behaviors, such as belittling or manipulation, which historically precede physical abuse. Preliminary results indicate that Multi-task Learning can improve detection of emotional abuse and help tune detection models to particular kinds of relationship abuse.

Shubhanshu Rao, Gaurav Kumar, Martin Agelin-Chaab

Achieving optimal design is a crucial aspect of any design process for safe and efficient operation. Such tasks typically require numerous simulations over many iterations, which can become computationally expensive. This paper proposes a novel method that combines Physics-informed Neural Networks (PINNs) with a Genetic Algorithm to optimize the parameters of an airfoil that aims to achieve favourable aerodynamic conditions. Traditional solvers are computationally expensive for performing such tasks, but using PINNs can significantly reduce this while keeping accuracy high. The proposed approach shows the advantage of using PINNs in optimizing complex engineering problems.

Dayu Qin, Yi Yan, Ercan Engin Kuruoglu

In this paper, we propose a novel framework that leverages Large Language Models (LLMs) for predicting missing values in time-varying graph signals by exploiting spatial and temporal smoothness. We leverage the power of LLM to achieve a message-passing scheme. For each missing node, its neighbors and previous estimates are fed into and processed by LLM to infer the missing observations. Tested on the task of the online prediction of wind-speed graph signals, our model outperforms online graph filtering algorithms in terms of accuracy, demonstrating the potential of LLMs in effectively addressing partially observed signals in graphs.

Howard Prioleau, Saurav Aryal

Adverse Drug Events (ADEs) are a major healthcare issue in the United States, contributing to millions of outpatient and emergency department visits and ranking as the fourth leading cause of death. While many ADEs are identified post-market, improved detection methods are crucial for enhancing patient safety. This study explores the application of large language models (LLMs) to the n2c2 task for ADE detection, evaluating optimal prompting techniques without requiring ADE-specific training data. Results indicate that an entity-only extraction approach outperforms the inline method, offering higher precision, recall, and token efficiency. This study highlights the potential of LLMs for accurate ADE detection in clinical text, improving performance while maintaining model efficiency.

Ruth-Emely Pierau, Alaster Meehan, Hamid Rezatofighi, Peter J. Stuckey

This abstract presents a simulated annealing based approach that constructs hyper-spectral images from the frequency spectrums of a distributed acoustic sensing system and iteratively improves them through the training of learnable filters. The aim is to construct an image that represents features of signals from events while repressing noise. Hyper-spectral images are specifically created for downstream computer vision tasks such as object detection. Hyper-spectral images are images with more than three channels that are derived from a frequency spectrum to obtain the spectrum for each image pixel. Simulated annealing is used to train the filters to automatically select frequencies and bin them into frequency bands. Each frequency band is mapped into an image channel. We fully integrate our filtering method with an object detection network so that filters are trained in conjunction with the neural network. The detection model serves as both the measure and the selector. Our simulated annealing approach significantly outperforms current state-of-the-art methods by a margin of 22%. Limitations include a dependency on randomness and excluding parts of the search space prematuraly due to the design of the local moves.

Ian Tong Pan, Joseph D. Romano

Molecular machine learning has broad applications across multiple domains such as drug development, environmental toxicology, and materials science. Various pre-trained frameworks using self-supervised representation learning have emerged to tackle the difficulty of obtaining large molecular datasets useful for training high-performing molecular machine learning models. In this study, we explore a novel representation learning framework trained using both 2D and 3D molecular data. Specifically, a 3D invariant graph neural network to learn how to capture 3D atomic information and then pass these atomic representations into a regular 2D graph neural network which can leverage molecular topology. Results from experiments demonstrate the representations produced by our method using both 3D and 2D molecular information lead to strong performance in downstream tasks.

Christianah Titilope Oyewale, Rafiat M. Bamimore Akodu, Faith Adeoluwa Adeyemi

Assessing social cognition in adolescents by understanding emotional perception is crucial, especially through tasks like the Reading the Mind in the Eyes Test (RMET). This ongoing research investigates the emotional perception skills of Nigerian high school girls through the Reading the Mind in the Eyes Test (RMET). Preliminary analysis looks at how age and class level (SS1 and SS2) affect RMET scores, with 20% of the data (n = 215) already gathered. ANOVA findings display a notable distinction among class levels (p = 0.024), while regression analysis suggests that age effectively predicts RMET scores (β = 1.06, p = 0.037), showcasing that older students achieve higher scores. These preliminary results indicate that age is a significant factor in how emotions are perceived, and more data is being collected and analyzed to gain deeper understanding. These findings can guide strategies to enhance social skills in education and improve AI models for emotion recognition in diverse and age-sensitive contexts.

Rongxin Ouyang, Kokil Jaidka, Subhayan Mukerjee, Guangyu Cui

The prevalence of multi-modal content on social media complicates automated moderation strategies. This calls for an enhancement in multi-modal classification and a deeper understanding of understated meanings in images and memes. Although previous efforts have aimed at improving model performance through fine-tuning, few have explored an end-to-end optimization pipeline that accounts for modalities, prompting, labelling, and fine-tuning. In this study, we propose an end-to-end conceptual framework for model opti- mization in complex tasks. Experiments support the efficacy of this traditional yet novel framework, achieving the highest accuracy and AUROC. Ablation experiments demonstrate that isolated optimisations are not ineffective on their own.

James T. Oswald, Brandon Rozek, Thomas Macaulay Ferguson, Selmer Bringsjord

We present our work on a new modal logic of optimality, OPT, whose semantics are modeled in terms of optimal paths through reward-weighted transition systems. We prove some basic properties of OPT, including its status as a normal modal logic, as well as its relation to some of the standard modal axioms. We end with a discussion of applications to AI and future research directions and extensions.

Ryusei Ohtani, Yuko Sakurai, Satoshi Oyama

Counterfactual explanations in Explainable AI (XAI) identify which features to change to alter an outcome, but existing methods adjust only the features of a single agent. We present a new approach to re-evaluating rankings that is based on predictions of future features of the other agents in a ranking system. It uses an algorithm that provides a more realistic counterfactual explanation of changing the ranking of a particular agent. Computer experiments demonstrated that the proposed algorithm can capture the time variation of the entire ranking system in the inference results.