Search

We present a novel and flexible learning-based method for generating tileable image sets. Our method goes beyond simple self-tiling, supporting sets of mutually tileable images that exhibit a high degree of diversity. To promote diversity we decouple structure from content by foregoing explicit copying of patches from an exemplar image. Instead we leverage the prior knowledge of natural images and textures embedded in large-scale pretrained diffusion models to guide tile generation constrained by exterior boundary conditions and a text prompt to specify the content. By carefully designing and selecting the exterior boundary conditions, we can reformulate the tile generation process as an inpainting problem, allowing us to directly employ existing diffusion-based inpainting models without the need to retrain a model on a custom training set. We demonstrate the flexibility and efficacy of our content-aware tile generation method on different tiling schemes, such as Wang tiles, from only a text prompt. Furthermore, we introduce a novel Dual Wang tiling scheme that provides greater texture continuity and diversity than existing Wang tile variants. [ABSTRACT FROM AUTHOR]

Large language models have demonstrated impressive capabilities in many domains. But they sometimes generate irrelevant or nonsensical text, or produce outputs that deviate from the provided input, an occurrence commonly referred to as hallucination. To mitigate this issue, we introduce a novel decoding method that incorporates both factual and hallucination prompts (DFHP). It applies contrastive decoding to highlight the disparity in output probabilities between factual prompts and hallucination prompts. Experiments on both multiple-choice and text generation tasks show that our approach significantly improves factual accuracy of large language models without additional training. On the TruthfulQA dataset, the DFHP method significantly improves factual accuracy of the LLaMA model, with an average improvement of 6.4% for the 7B, 13B, 30B, and 65B versions. Its high accuracy in factuality makes it an ideal choice for high reliability tasks like medical diagnosis and legal cases. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Economics & Administrative Sciences / Afyon Kocatepe Üniversitesi Iktisadi ve Idari Bilimler Fakültesi Dergisi is the property of Afyon Kocatepe University, Faculty of Business Administration and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Artificial Intelligence (AI) plays an increasingly prominent role in various spheres of life in today’s world, including generation of a variety of visual content from selfie stream processing to creating works of digital art. The present paper raises the question of whether AI is capable of creating real art or it just imitates its external form. The paper examines the specificity of prompts: from concrete named ones to interpretive descriptive queries in linguistic, artistic and socio-cultural contexts. The article dwells upon some important aspects of evaluating the quality of keyword extraction algorithms and their relation to artistic practice. The authors rely on semiotic analysis to uncover encoded meanings and imports in the text. The article emphasises that the literary text is at the top of the hierarchy of cultural texts; it is characterised by intentionality and coherence and represents a complex semantic field where key words and images interact with the explicit and implicit contexts. The study examines and analyses the visualised images of Cheshire Cat, Cat Behemoth and Tomcat Murr created by the authors with the use of three generative neural networks: Stable Diffusion, Dall‑E and Kandinsky. Understanding and visualising the literary text by generative systems and models realising specific algorithms requires the ability to reveal its multilayered semantics and connection with the cultural context, which ultimately helps to understand the in-depth meanings of the work and its place in culture. Consideration of the operational quality of algorithms for keyword system extraction and image generation is deemed possible from the point of view of their structural organisation. Generative algorithms create an imitative reality, while the immanence of the artistic value determines the uniqueness and meanings of the created figurative world. The article can be useful to anyone interested in the substance and specificity of digital art, the relationship between technological innovations and socio-cultural context, the creation and visualisation of artistic images in generative AI models, their conceptualisation and interpretation.

The intelligent judicial decision-making is gradually aligning with the logic of legal adjudication. Case element recognition is a fundamental task proposed in recent years. Compared with earlier methods based on deep learning and machine reading comprehension, the generative element recognition approach using large language models (LLM) holds greater potential for complex reasoning. However, the current performance of judicial LLM on these fundamental tasks remains suboptimal. This paper introduces a conversational mixture of expert element recognition LLM. The proposed model in this paper first designs specific prompts tailored to the characteristics of cases for the ChatGLM3-6B-base model. The LLM is then fine-tuned with full parameters to acquire basic element recognition capabilities, with its weights shared among subsequent hybrid experts to reduce learning costs. To address different case types and label imbalance scenarios, case-specific DoRA experts and label-specific DoRA experts are integrated into the LLM’s attention layer, enhancing the model’s ability to differentiate between tasks. A learnable gating mechanism is also designed to facilitate the selection of label experts. The proposed model is tested on the CAIL2019 dataset and a desensitized theft case element recognition dataset from a certain province, nine benchmark models across three types of methods are compared, and ablation experiments are conducted. Experimental results show that the proposed model’s overall performance, measured by the F1 score, exceeds the best-performance model by 5.9 percentage points. On the label-imbalanced CAIL2019 dataset, the label expert effectively mitigates the impact of extreme data imbalance. Additionally, without repeated full-parameter fine-tuning, the basic model trained on CAIL2019 achieves optimal results in theft cases of a certain province after lightweight fine-tuning by case and label experts, demonstrating the model’s scalability.

Kamu Denetçiliği Kurumu, daha yaygın bilinen adıyla Ombudsmanlık, toplumda şeffaflık ve adalet ilkesinin korunması adına denetim mekanizmasında önemli rol üstlenen bir kurum olarak değerlendirilmektedir. Bu kurum, vatandaşların kamu hizmetlerinden adil bir şekilde yararlanmasını sağlamak için önemli bir denetim aracı olarak işlev görmektedir. Özellikle, kamu personel rejimi konusunda aldığı tavsiye kararlarıyla dikkat çekmektedir. Bu kararlar, kamu kurumlarının personel politikalarını denetleyerek adaletin sağlanmasına katkıda bulunmaktadır. Kurumun yeni nesil teknolojilerden faydalanması, karar alma süreçlerinin etkinliğini artırabileceği sorusu akla gelmektedir. Örneğin, yapay zekâ teknolojisi, büyük veri setlerini hızlı ve objektif bir şekilde analiz ederek karar alma süreçlerine katkı sağlayabilir. Yapay zekâ tabanlı uygulamalar arasında yer alan ChatGPT de bu tavsiye kararlarının değerlendirilmesinde önemli bir araç olabilir. Bu teknolojilerin kullanımı, kurumun kararlarını daha kapsamlı ve etkili bir şekilde değerlendirmesine olanak tanıyabilir. Kamu Denetçiliği Kurumu tarafından alınan tavsiye kararlarının değerlendirilmesinde, ChatGPT gibi yapay zekâ tabanlı geliştirilen uygulamaların kullanımının kurum tarafından alınan tavsiye kararlarının etkinliğini ve verimliliğini artırabilir. Bu çalışma da yapay zekâ teknolojilerinden biri olan ChatGPT’nin Kamu Denetçiliği Kurumunun kararlarının değerlendirilmesindeki potansiyelini incelemeyi ve bu teknolojinin potansiyelini tartışmayı amaçlamaktadır.

Abstract Aspect-based sentiment analysis (ABSA) is a challenging task due to the presence of multiple aspect words with different sentiment polarities in a sentence. Recently, pre-trained language models like BERT have been widely used as context encoders in ABSA. Graph neural networks have also been employed to extract syntactic and semantic information from sentence parsing trees, resulting in superior results. However, dependency trees may establish irrelevant dependencies for sentences with irregular syntax and complex structures. Additionally, previous methods have not fully utilized recent developments in pre-trained language models. Therefore, we propose a Dual Syntax aware Graph attention networks with Prompt (DSGP) model to address these issues. Our model utilizes prompt templates to maximize the potential of pre-trained models and masked vector outputs of templates as supplementary aspect feature representations. We also leverage both dependency trees and constituent trees with graph attention networks to extract different types of syntactic information. The dependency tree captures syntactic correlation between words, while the constituent tree provides a high-level formation of the sentence. Finally, the output from the prompt and parsing trees is fused and fed into a standard classifier. Experimental results on four public datasets demonstrate the competitive performance of our model.

Copyright of Acta Comportamentalia is the property of Instituto de Psicologia y Educacion de la Universidad Veracruzana and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Aspect Sentiment Quad Prediction is a research topic of paramount significance and complexity within the Aspect-Based Sentiment Analysis task. Leveraging the generative paradigm of the T5 model, we achieve end-to-end extraction of aspect sentiment elements by paraphrasing the original text into sentences predefined by templates. Current research predominantly confines templates to single sentences or directly concatenates sentiment elements using a few symbols, limiting the model's reasoning opportunities. In this work, we introduce a Self-Inference Template (SIT) to guide the model in thoughtful reasoning, facilitating a step-by-step inference generation process. This approach enables the model to more accurately identify aspect sentiment elements and their interdependencies. Experimental results demonstrate a significant improvement in quadruplet prediction performance under constant time costs, effectively mitigating overfitting issues caused by limited data volume to some extent. [ABSTRACT FROM AUTHOR]

The rapid advancements in artificial intelligence (AI) have transformed various domains, including education. Generative AI models have garnered significant attention for their potential in educational settings, but image-generative AI models need to be more utilized. This study explores the potential of integrating generative AI, specifically Stable Diffusion, into art-focused STEAM classes. To this end, we combined STEAM education with image-generative AI. We designed and implemented a learning activity in which students used image-generative AI to generate creative images and wrote imaginative diaries inspired by AI-generated pictures. Using a mixed-methods approach, we collected and analyzed data from surveys, interviews, and prompt datasets to gain insights into students' perceptions, motivations, and behaviors concerning the integration of generative AI in art education. Our study found that incorporating generative AI into art-focused STEAM classes may reduce the gap between male and female students' interest in art practices. We also found that the number of subjects in students' prompt datasets is linked to their divergent and convergent thinking. Finally, we discovered a meaningful correlation between the number of unique words in students' diaries and the image scores that researchers assessed. Our findings shed light on the relationship between students' divergent and convergent behaviors, offering valuable insights into their creative thinking processes. This study underscores the potential of integrating image-generative AI models in STEAM education. [ABSTRACT FROM AUTHOR]

Las aplicaciones de inteligencia artificial generativa (IAG) como ChatGPT, Gemini o DeepL, están transformando de manera profunda las prácticas letradas de la comunidad, al presentarse como asistentes eficaces de lectores y escritores. En este marco, exponemos una visión de conjunto con los beneficios potenciales y los riesgos éticos que supone incorporar la IAG en la educación media y superior. Explicamos algunos conceptos básicos del ámbito, como inteligencia artificial (generativa), modelo de lenguaje o algoritmo, además de esbozar la línea cronológica de desarrollo hasta la actualidad de esta tecnología. Documentamos diversos ejemplos empíricos de prácticas letradas expertas de varios profesionales (periodista, abogado, publicista). A continuación, abordamos las potencialidades que ofrece la IAG para el docente y para el aprendiz; presentamos un protocolo didáctico para legitimarla en el aula, con énfasis en la importancia del aprendizaje autorregulado o adaptativo, y cerramos con varias reflexiones sobre la época del posplagio y la escritura híbrida humano/IAG.

The combination of prompt learning and contrastive learning has recently been a promising approach to few-shot learning in NLP field. However, most of these studies only focus on the semantic-level relevance and intra-class information of data in the class level while ignoring the importance of fine-grained instance-level feature representations. This paper proposes a joint contrastive learning (JCL) framework that leverages instance-level contrastive learning to learn fine-grained differences of feature representations and class-level contrastive learning to learn richer intra-class information. The experimental results demonstrate that the proposed JCL method is effective and has strong generalization ability. Our code is available at https://github.com/2251821381/JCL. [ABSTRACT FROM AUTHOR]

The rapid growth of fake news on the Internet poses a challenge to accessing authentic information. Fake news detection plays a crucial role in filtering out false information and improving information accuracy. However, practical implementation faces challenges like high annotation costs, limited samples, poor training results, and weak model generalization. To tackle these issues, we propose knowledge-enhanced AutoPrompt (KEAP) for fake news detection. This method leverages prompt templates generated by the T5 model to transform the fake news detection task into a prompt learning-based task. We also incorporate external entity knowledge to enhance detection capabilities. Carefully designed prompts activate the model's latent knowledge, improving performance in low-resource scenarios and model generalization. Experiments on GossipCop and PolitiFact datasets demonstrate the superiority of prompt learning over existing methods without extra text during testing. KEAP achieves an average F1 score improvement of 2.79 % compared to state-of-the-art methods in few-shot settings. [ABSTRACT FROM AUTHOR]

Artificial intelligence (AI), focused on systems capable of emulating complex human tasks, has experienced a notable boom with the opening of platforms such as ChatGPT. However, these models, specialized in text generation, lack specific training for scientific research, which can compromise their reliability and ability to be truly useful in research work. Key strategies to optimize interaction with AI are explored, highlighting the importance of prompts. In addition, tools specifically designed for scientific research are presented. Ethical and privacy considerations when using AI are also addressed, emphasizing that these tools do not replace traditional methods, but rather complement research work. Finally, the role of AI in data analysis, assisted writing and document management is highlighted, underlining the need to learn to take advantage of these emerging technologies to improve efficiency in scientific research. [ABSTRACT FROM AUTHOR]

This paper examines the ability to use Stable Diffusion's diffusion models to get state-of-the-art synthesis results on image data and other types of data. Also, a guiding interface can be used to control the process of making images by converting text to images and image to image. But because these models usually work directly in pixel space, optimizing strong DMs often needs more GPU VRAM to run. Using Stable Diffusion and diffusion models on local hardware like this lets more information and depth be added while generating images, which greatly improves the quality detail of the image. By combining diffusion models to model architecture, I have made diffusion models into powerful and flexible producers for general conditioning inputs, such as when using XL-XDXL 1.0 and LoRA models. Overall, the paper highlights how a normal person can run their own Midjourney like AI image generation with the help of machine learning and generative AI.

Large language models have demonstrated impressive capabilities in many domains. But they sometimes generate irrelevant or nonsensical text, or produce outputs that deviate from the provided input, an occurrence commonly referred to as hallucination. To mitigate this issue, we introduce a novel decoding method that incorporates both factual and hallucination prompts (DFHP). It applies contrastive decoding to highlight the disparity in output probabilities between factual prompts and hallucination prompts. Experiments on both multiple-choice and text generation tasks show that our approach significantly improves factual accuracy of large language models without additional training. On the TruthfulQA dataset, the DFHP method significantly improves factual accuracy of the LLaMA model, with an average improvement of 6.4% for the 7B, 13B, 30B, and 65B versions. Its high accuracy in factuality makes it an ideal choice for high reliability tasks like medical diagnosis and legal cases.

Diffusion models have achieved tremendous success in modeling continuous data modalities, such as images, audio, and video, yet their application in discrete data domains (e.g., natural language) has been limited. Existing methods primarily represent discrete text in a continuous diffusion space, incurring significant computational overhead during training and resulting in slow sampling speeds. This paper introduces LaDiffuSeq, a latent diffusion-based text generation model incorporating an encoder–decoder structure. Specifically, it first employs a pretrained encoder to map sequences composed of attributes and corresponding text into a low-dimensional latent vector space. Then, without the guidance of a classifier, it performs the diffusion process for the sequence's corresponding latent space. Finally, a pretrained decoder is used to decode the newly generated latent vectors, producing target texts that are relevant to themes and possess multiple emotional granularities. Compared to the benchmark model, DiffuSeq, this model achieves BERTScore improvements of 0.105 and 0.009 on two public real-world datasets (ChnSentiCorp and a debate dataset), respectively; perplexity falls by 3.333 and 4.562; and it effectively quadruples the text generation sampling speed. [ABSTRACT FROM AUTHOR]