Evaluating the Performance of ChatGPT in Ophthalmology

Ophthalmology Science, Год журнала: 2023, Номер 3(4), С. 100324 - 100324

Опубликована: Май 5, 2023

Foundation models are a novel type of artificial intelligence algorithms, in which pretrained at scale on unannotated data and fine-tuned for myriad downstream tasks, such as generating text. This study assessed the accuracy ChatGPT, large language model (LLM), ophthalmology question-answering space.

Язык: Английский

---

Generative AI

Business & Information Systems Engineering, Год журнала: 2023, Номер 66(1), С. 111 - 126

Опубликована: Сен. 12, 2023

The term "generative AI" refers to computational techniques that are capable of generating seemingly new, meaningful content such as text, images, or audio from training data. widespread diffusion this technology with examples Dall-E 2, GPT-4, and Copilot is currently revolutionizing the way we work communicate each other. In article, provide a conceptualization generative AI an entity in socio-technical systems models, systems, applications. Based on that, introduce limitations current agenda for Business & Information Systems Engineering (BISE) research. Different previous works, focus context information and, end, discuss several opportunities challenges unique BISE community make suggestions impactful directions

Язык: Английский

---

Florence: A New Foundation Model for Computer Vision

arXiv (Cornell University), Год журнала: 2021, Номер unknown

Опубликована: Янв. 1, 2021

Automated visual understanding of our diverse and open world demands computer vision models to generalize well with minimal customization for specific tasks, similar human vision. Computer foundation models, which are trained on diverse, large-scale dataset can be adapted a wide range downstream critical this mission solve real-world applications. While existing such as CLIP, ALIGN, Wu Dao 2.0 focus mainly mapping images textual representations cross-modal shared representation, we introduce new model, Florence, expand the from coarse (scene) fine (object), static (images) dynamic (videos), RGB multiple modalities (caption, depth). By incorporating universal visual-language Web-scale image-text data, Florence model easily various classification, retrieval, object detection, VQA, image caption, video retrieval action recognition. Moreover, demonstrates outstanding performance in many types transfer learning: fully sampled fine-tuning, linear probing, few-shot zero-shot novel objects. All these properties serve general purpose tasks. achieves state-of-the-art results majority 44 representative benchmarks, e.g., ImageNet-1K classification top-1 accuracy 83.74 top-5 97.18, 62.4 mAP COCO tuning, 80.36 87.8 Kinetics-600.

Язык: Английский

---

Towards a standard for identifying and managing bias in artificial intelligence

Опубликована: Март 15, 2022

As individuals and communities interact in with an environment that is increasingly virtual they are often vulnerable to the commodification of their digital exhaust. Concepts behavior ambiguous nature captured this environment, quantified, used categorize, sort, recommend, or make decisions about people's lives. While many organizations seek utilize information a responsible manner, biases remain endemic across technology processes can lead harmful impacts regardless intent. These outcomes, even if inadvertent, create significant challenges for cultivating public trust artificial intelligence (AI). SP 1270 NIST Artificial Intelligence publication should be read conjunction all publications AI Series, which was established January 2023.

Язык: Английский

---

FLAVA: A Foundational Language And Vision Alignment Model

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Год журнала: 2022, Номер unknown, С. 15617 - 15629

Опубликована: Июнь 1, 2022

State-of-the-art vision and vision-and-language models rely on large-scale visio-linguistic pretraining for obtaining good performance a variety of downstream tasks. Generally, such are often either cross-modal (contrastive) or multi-modal (with earlier fusion) but not both; they only target specific modalities A promising direction would be to use single holistic universal model, as "foundation", that targets all at once-a true language foundation model should tasks, cross- We introduce FLAVA demonstrate impressive wide range 35 tasks spanning these modalities.

Язык: Английский

---

Chatting and Cheating. Ensuring academic integrity in the era of ChatGPT

Опубликована: Янв. 10, 2023

The use of artificial intelligence in academia is a hot topic the education field. chatAPIs and GPT-3 higher has potential to offer range benefits, including increased student engagement, collaboration, accessibility. However, these tools also raise number challenges concerns, particularly relation academic honesty plagiarism. This paper examines opportunities using education, with focus on risks rewards ways which universities can address they pose. discusses main features capabilities provides examples their education. It considers for be used dishonesty difficulties detecting preventing such abuses. Finally, suggests strategies that adopt ensure are ethically responsibly, developing policies procedures, providing training support, variety methods detect prevent cheating.

Язык: Английский

---

scGPT: toward building a foundation model for single-cell multi-omics using generative AI

Nature Methods, Год журнала: 2024, Номер 21(8), С. 1470 - 1480

Опубликована: Фев. 26, 2024

Язык: Английский

---

SpectralGPT: Spectral Remote Sensing Foundation Model

IEEE Transactions on Pattern Analysis and Machine Intelligence, Год журнала: 2024, Номер 46(8), С. 5227 - 5244

Опубликована: Апрель 3, 2024

The foundation model has recently garnered significant attention due to its potential revolutionize the field of visual representation learning in a self-supervised manner. While most models are tailored effectively process RGB images for various tasks, there is noticeable gap research focused on spectral data, which offers valuable information scene understanding, especially remote sensing (RS) applications. To fill this gap, we created first time universal RS model, named SpectralGPT, purpose-built handle using novel 3D generative pretrained transformer (GPT). Compared existing models, SpectralGPT 1) accommodates input with varying sizes, resolutions, series, and regions progressive training fashion, enabling full utilization extensive Big Data; 2) leverages token generation spatial-spectral coupling; 3) captures spectrally sequential patterns via multi-target reconstruction; 4) trains one million images, yielding over 600 parameters. Our evaluation highlights performance improvements signifying substantial advancing Data applications within geoscience across four downstream tasks: single/multi-label classification, semantic segmentation, change detection.

Язык: Английский

---

AI Chains: Transparent and Controllable Human-AI Interaction by Chaining Large Language Model Prompts

CHI Conference on Human Factors in Computing Systems, Год журнала: 2022, Номер unknown, С. 1 - 22

Опубликована: Апрель 28, 2022

Although large language models (LLMs) have demonstrated impressive potential on simple tasks, their breadth of scope, lack transparency, and insufficient controllability can make them less effective when assisting humans more complex tasks. In response, we introduce the concept Chaining LLM steps together, where output one step becomes input for next, thus aggregating gains per step. We first define a set primitive operations useful Chain construction, then present an interactive system users modify these Chains, along with intermediate results, in modular way. 20-person user study, found that not only improved quality task outcomes, but also significantly enhanced controllability, sense collaboration. Additionally, saw developed new ways interacting LLMs through Chains: they leveraged sub-tasks to calibrate model expectations, compared contrasted alternative strategies by observing parallel downstream effects, debugged unexpected outputs "unit-testing" sub-components Chain. two case studies, further explore how Chains may be used future applications.

Язык: Английский

---

Image Segmentation Using Text and Image Prompts

2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Год журнала: 2022, Номер unknown, С. 7076 - 7086

Опубликована: Июнь 1, 2022

Image segmentation is usually addressed by training a model for fixed set of object classes. Incorporating additional classes or more complex queries later expensive as it requires re-training the on dataset that encompasses these expressions. Here we propose system can generate image segmentations based arbitrary prompts at test time. A prompt be either text an image. This approach enables us to create unified (trained once) three common tasks, which come with distinct challenges: referring expression segmentation, zero-shot and one-shot segmentation. We build upon CLIP backbone extend transformer-based decoder dense prediction. After extended version PhraseCut dataset, our generates binary map free-text expressing query. analyze different variants latter image-based in detail. novel hybrid input allows dynamic adaptation not only tasks mentioned above, but any task where query formulated. Finally, find adapt well generalized involving affordances properties. Code available https://eckerlab.org/code/CLIPSeg

Язык: Английский

---