论文检索

Shixuan Liu, Chao Wang, Shuangyong Song

Pre-trained language models have shown their high performance of text processing in intelligent customer service platforms. However, these models do not leverage domain specific information. In this paper, we propose icsPLMs optimized for intelligent customer service on both word and sentence levels. Our experimental results represent that using targeted strategies can further improve the performance of pre-trained language models in this field.

Hanyue Liu, Haonan Cheng, Long Ye

Existing image-to-image translation methods perform less satisfactorily in the "day-night" domain due to insufficient scene feature study. To address this problem, we propose DNIT, which performs fine-grained handling of features by a nighttime image preprocessing (NIP) module and an edge fusion detection (EFD) module. The NIP module enhances brightness while minimizing noise, facilitating extracting content and style features. Meanwhile, the EFD module utilizes two types of edge images as additional constraints to optimize the generator. Experimental results show that we can generate more realistic and higher-quality images compared to other methods, proving the effectiveness of our DNIT.

Jiaxin Liang, Junping Zhou, Minghao Yin

The Diversified Top-k MaxSAT (DTKMS) problem is an extension of MaxSAT. The objective of DTKMS is to find k feasible assignments of a given formula, such that each assignment satisfies all hard clauses and the k assignments together satisfy the maximum number of soft clauses. This paper presents a local search algorithm, DTKMS-DIA, which incorporates a new approach to generating initial assignments. Experimental results indicate that DTKMS-DIA can achieve attractive performance on 826 instances compared with state-of-the-art solvers.

Zhengyu Li, Curtis Bright, Vijay Ganesh

The problem of finding the minimum three-dimensional Kochen–Specker (KS) vector system, an important problem in quantum foundations, has remained open for over 55 years. We present a new method to address this problem based on a combination of a Boolean satisfiability (SAT) solver and a computer algebra system (CAS). Our approach improved the lower bound on the size of a KS system from 22 to 24. More importantly, we provide the first computer-verifiable proof certificate of a lower bound to the KS problem with a proof size of 41.6 TiB for order 23. The efficiency is due to the powerful combination of SAT solvers and CAS-based orderly generation.

Yue Lei, Bin Chen, Wenxin Tai, Ting Zhong, Fan Zhou

Recently, the field of Speech Enhancement has witnessed the success of diffusion-based generative models. However, these diffusion-based methods used to take multiple iterations to generate high-quality samples, leading to high computational costs and inefficiency. In this paper, we propose SDFEN (Shallow Diffusion for Fast spEech eNhancement), a novel approach for addressing the inefficiency problem while enhancing the quality of generated samples by reducing the iterative steps in the reverse process of diffusion method. Specifically, we introduce the shallow diffusion strategy initiating the reverse process with an adaptive time step to accelerate inference. In addition, a dedicated noisy predictor is further proposed to guide the adaptive selection of time step. Experiment results demonstrate the superiority of the proposed SDFEN in effectiveness and efficiency.

Mu-Tien Kuo, Chih-Chung Hsueh, Richard Tzong-Han Tsai

As Large Language Models (LLMs) become more prevalent in various fields, it is crucial to rigorously assess the quality of their explanations. Our research introduces a task-agnostic framework for evaluating free-text rationales, drawing on insights from both linguistics and machine learning. We evaluate two dimensions of explainability: fidelity and interpretability. For fidelity, we propose methods suitable for proprietary LLMs where direct introspection of internal features is unattainable. For interpretability, we use language models instead of human evaluators, addressing concerns about subjectivity and scalability in evaluations. We apply our framework to evaluate GPT-3.5 and the impact of prompts on the quality of its explanations. In conclusion, our framework streamlines the evaluation of explanations from LLMs, promoting the development of safer models.

Shivanand Kundargi, Tejas Anvekar, Ramesh Tabib, Uma Mudenagudi

Neural Radiance Fields (NeRF) have been extensively explored as a leading approach for modeling and representing 3D data across various domains. Their ability to capture arbitrary scale point clouds and generate novel views makes them particularly valuable for digitizing cultural heritage sites. However, despite their impressive rendering capabilities, prior methods have often overlooked a significant real-world challenge: handling open-world scenarios characterized by unstructured data containing multiple classes in a single set of unlabeled images. To address this challenge, we propose a novel method NCD-NeRF that leverages Novel-Class Discovery to effectively tackle the complexities inherent in real-world data with unlabeled classes while excelling in producing high-quality NeRF representation. To validate our approach, we conducted a benchmarking analysis using a custom-collected dataset featuring UNESCO World Heritage sites in India. We observe that our proposed NCD-NeRF can parallely discover novel classes and render high-quality 3D volumes.

Daya Kumar, Abhijith Sharma, Apurva Narayan

Convolutional neural networks (CNNs) are being increasingly adopted in medical imaging. However, in the race for developing accurate models, their robustness is often overlooked. This elicits a significant concern given the safety-critical nature of the healthcare system. Here, we highlight the vulnerability of CNNs against a sporadic and naturalistic adversarial patch attack (SNAP). We train SNAP to mislead the ResNet50 model predicting metastasis in histopathological scans of lymph node sections, lowering the accuracy by 27%. This work emphasizes the need for defense strategies before deploying CNNs in critical healthcare settings.

Ryan Koo, Yekyung Kim, Dongyeop Kang, Jaehyung Kim

Detecting out-of-distribution (OOD) samples is crucial for robust NLP models. Recent works observe two OOD types: background shifts (style change) and semantic shifts (content change), but existing detection methods vary in effectiveness for each type. To this end, we propose Meta-Crafting, a unified OOD detection method by constructing a new discriminative feature space utilizing 7 model-driven metadata chosen empirically that well detects both types of shifts. Our experimental results demonstrate state-of-the-art robustness to both shifts and significantly improved detection on stress datasets.

Rui Kong, Chenyang Wu, Zongzhang Zhang

Current policy gradient techniques excel in refining policies over sampled states but falter when generalizing to unseen states. To address this, we introduce Reinforcement Sampling (RS), a novel method leveraging a generalizable action value function to sample improved decisions. RS is able to improve the decision quality whenever the action value estimation is accurate. It works by improving the agent's decision on the fly on the states the agent is visiting. Compared with the historically experienced states in which conventional policy gradient methods improve the policy, the currently visited states are more relevant to the agent. Our method sufficiently exploits the generalizability of the value function on unseen states and sheds new light on the future development of generalizable reinforcement learning.

Taeyoung Kim, Dongsoo Har

In multi-goal reinforcement learning with a sparse binary reward, training agents is particularly challenging, due to a lack of successful experiences. To solve this problem, hindsight experience replay (HER) generates successful experiences even from unsuccessful ones. However, generating successful experiences from uniformly sampled ones is not an efficient process. In this paper, the impact of exploiting the property of achieved goals in generating successful experiences is investigated and a novel cluster-based sampling strategy is proposed. The proposed sampling strategy groups episodes with different achieved goals by using a cluster model and samples experiences in the manner of HER to create the training batch. The proposed method is validated by experiments with three robotic control tasks of the OpenAI Gym. The results of experiments demonstrate that the proposed method is substantially sample efficient and achieves better performance than baseline approaches.

Jongseok Kim, Ohyun Jo

This work proposes and analyzes IncepSeqNet which is a new model combining the Inception Module with the innovative Multi-Shape Augmentation technique. IncepSeqNet excels in feature extraction from sequence signal data consisting of a number of complex numbers to achieve superior classification accuracy across various SNR(Signal-to-Noise Ratio) environments. Experimental results demonstrate IncepSeqNet’s outperformance of existing models, particularly at low SNR levels. Furthermore, we have confirmed its applicability in practical 5G systems by using real-world signal data.

Aditya Kasliwal, Aryan Kamani, Ishaan Gakhar, Pratinav Seth, Sriya Rallabandi

Recent advances in image-to-image translation involve the integration of non-visual imagery in deep models. Non-visual sensors, although more costly, often produce low-resolution images. To combat this, methods using RGB images to enhance the resolution of these modalities have been introduced. Fusing these modalities to achieve high-resolution results demands models with millions of parameters and extended inference times. We present LaMAR, a lightweight model. It employs Laplacian image pyramids combined with a low-resolution thermal image for Guided Thermal Super Resolution. By decomposing the RGB image into a Laplacian pyramid, LaMAR preserves image details and avoids high-resolution feature map computations, ensuring efficiency. With faster inference times and fewer parameters, our model demonstrates state-of-the-art results.

Yash Kankariya, Suguman Bansal

Prior compositional methods in LTLf to DFA conversion have focussed on improving the composition phase. In this work, we examine improvements to the decomposition phase that result in overall improvements in LTLf to DFA translation. Our work is based on reducing the structure of the underlying Abstract Syntax Tree (AST) of a formula such that the new AST results in fewer composition operations.

Seung Woo Kang, Ohyun Jo

We present an imagification approach for multivariate time-series data tailored to constrained NN-based forecasting model training environments. Our imagification process consists of two key steps: Re-stacking and time embedding. In the Re-stacking stage, time-series data are arranged based on high correlation, forming the first image channel using a sliding window technique. The time embedding stage adds two additional image channels by incorporating real-time information. We evaluate our method by comparing it with three benchmark imagification techniques using a simple CNN-based model. Additionally, we conduct a comparison with LSTM, a conventional time-series forecasting model. Experimental results demonstrate that our proposed approach achieves three times faster model training termination while maintaining forecasting accuracy.

Inwon Kang, Parikshit Ram, Yi Zhou, Horst Samulowitz, Oshani Seneviratne

Data distillation is a technique of reducing a large dataset into a smaller dataset. The smaller dataset can then be used to train a model which can perform comparably to a model trained on the full dataset. Past works have examined this approach for image datasets, focusing on neural networks as target models. However, tabular datasets pose new challenges not seen in images. A sample in tabular dataset is a one dimensional vector unlike the two (or three) dimensional pixel grid of images, and Non-NN models such as XGBoost can often outperform neural network (NN) based models. Our contribution in this work is two-fold: 1) We show in our work that data distillation methods from images do not translate directly to tabular data; 2) We propose a new distillation method that consistently outperforms the baseline for multiple different models, including non-NN models such as XGBoost.

Eul Ka, Seungeun Go, Minjin Kwak, Jeong-Hun Kim, Aziz Nasridinov

Solar power generation has recently been in the spotlight as global warming continues to worsen. However, two significant problems may hinder solar power generation, considering that solar panels are installed outside. The first is soiling, which accumulates on solar panels, and the second is a decrease in sunlight owing to bad weather. In this paper, we will demonstrate that the solar power generation forecasting can increase when considering soiling and sunlight information. We first introduce a dataset containing images of clean and soiled solar panels, sky images, and weather information. For accurate solar power generation forecasting, we propose a new multimodal model that aggregates various features related to weather, soiling, and sunlight. The experimental results demonstrated the high accuracy of our proposed multimodal model.

Jinsun Jung, Hyeoneui Kim

Explainable Artificial Intelligence (XAI), a promising future technology in the field of healthcare, has attracted significant interest. Despite ongoing efforts in the development of XAI approaches, there has been inadequate evaluation of explanation effectiveness and no standardized framework for the evaluation has been established. This study aims to examine the relationship between subjective interpretability and perceived plausibility for various XAI explanations and to determine the factors affecting users' acceptance of the XAI explanation.

Amelia Jobe, Richard Ky, Sandra Luo, Akshay Dhamsania, Sumit Purohit, Edoardo Serra

Cyberattacks on power grids pose significant risks to national security. Power grid attacks typically lead to abnormal readings in power output, frequency, current, and voltage. Due to the interconnected structure of power grids, abnormalities can spread throughout the system and cause widespread power outages if not detected and dealt with promptly. Our research proposes a novel anomaly detection system for power grids that prevents overfitting. We created a network graph to represent the structure of the power grid, where nodes represent power grid components like generators and edges represent connections between nodes such as overhead power lines. We combine the capabilities of Long Short-Term Memory (LSTM) models with a Graph Isomorphism Network (GIN) in a hybrid model to pinpoint anomalies in the grid. We train our model on each category of nodes that serves a similar structural purpose to prevent overfitting of the model. We then assign each node in the graph a unique signature using a GIN. Our model achieved a 99.92% accuracy rate, which is significantly higher than a version of our model without structural encoding, which had an accuracy level of 97.30%. Our model allows us to capture structural and temporal components of power grids and develop an attack detection system with high accuracy without overfitting.

Yixuan Jin, Yutao Wei, Zhangtao Cheng, Wenxin Tai, Chunjing Xiao, Ting Zhong

The success of graph neural networks (GNNs) has spurred numerous new works leveraging GNNs for modeling multivariate time series anomaly detection. Despite their achieved performance improvements, most of them only consider static graph to describe the spatial-temporal dependencies between time series. Moreover, existing works neglect the time and scale-changing structures of time series. In this work, we propose MDGAD, a novel multi-scale dynamic graph structure learning approach for time series anomaly detection. We design a multi-scale graph structure learning module that captures the complex correlations among time series, constructing an evolving graph at each scale. Meanwhile, an anomaly detector is used to combine bilateral prediction errors to detect abnormal data. Experiments conducted on two time series datasets demonstrate the effectiveness of MDGAD.