Papers Read on AI

ChatGPT is attracting a cross-field interest as it provides a language interface with remarkable conversational competency and reasoning capabilities across many domains. However, since ChatGPT is trained with languages, it is currently not capable of processing or generating images from the visual world. At the same time, Visual Foundation Models, such as Visual Transformers or Stable Diffusion, although showing great visual understanding and generation capabilities, they are only experts on specific tasks with one-round fixed inputs and outputs. To this end, We build a system called Visual ChatGPT, incorporating different Visual Foundation Models, to enable the user to interact with ChatGPT

2023: Chenfei Wu, Sheng-Kai Yin, Weizhen Qi, Xiaodong Wang, Zecheng Tang, Nan Duan

https://arxiv.org/pdf/2303.04671v1.pdf

We introduce LLaMA, a collection of foundation language models ranging from 7B to 65B parameters. We train our models on trillions of tokens, and show that it is possible to train state-of-the-art models using publicly available datasets exclusively, without resorting to proprietary and inaccessible datasets. In particular, LLaMA-13B outperforms GPT-3 (175B) on most benchmarks, and LLaMA-65B is competitive with the best models, Chinchilla-70B and PaLM-540B. We release all our models to the research community.

2023: Hugo Touvron, Thibaut Lavril, Gautier Izacard, Xavier Martinet, Marie-Anne Lachaux, Timothée Lacroix, Baptiste Rozière, Naman Goyal, Eric Hambro, Faisal Azhar, Aur'elien Rodriguez, Armand Joulin, Edouard Grave, Guillaume Lample

Ranked #1 on Question Answering on PIQA

https://arxiv.org/pdf/2302.13971v1.pdf

Introduced by Hinton et al. in 2012, dropout has stood the test of time as a regularizer for preventing overfitting in neural networks. In this study, we demonstrate that dropout can also mitigate underfitting when used at the start of training. During the early phase, we find dropout reduces the directional variance of gradients across mini-batches and helps align the mini-batch gradients with the entire dataset's gradient. This helps counteract the stochasticity of SGD and limit the influence of individual batches on model training. Our findings lead us to a solution for improving performance in underfitting models.

2023: Zhuang Liu, Zhi Xu, Joseph Jin, Zhiqiang Shen, Trevor Darrell

https://arxiv.org/pdf/2303.01500v1.pdf

Diffusion models have enabled high-quality, conditional image editing capabilities. We propose to expand their arsenal, and demonstrate that off-the-shelf diffusion models can be used for a wide range of cross-domain compositing tasks. Among numerous others, these include image blending, object immersion, texture-replacement and even CG2Real translation or stylization. We employ a localized, iterative refinement scheme which infuses the injected objects with contextual information derived from the background scene, and enables control over the degree and types of changes the object may undergo. We conduct a range of qualitative and quantitative comparisons to prior work, and exhibit that our method produces higher quality and realistic results without requiring any annotations or training. Finally, we demonstrate how our method may be used for data augmentation of downstream tasks.

2023: Roy Hachnochi, Mingrui Zhao, Nadav Orzech, Rinon Gal, Ali Mahdavi-Amiri, D. Cohen-Or, Amit H. Bermano

https://arxiv.org/pdf/2302.10167v1.pdf

Tabular data is a common form of organizing data. Multiple models are available to generate synthetic tabular datasets where observations are independent, but few have the ability to produce relational datasets. Modeling relational data is challenging as it requires modeling both a"parent"table and its relationships across tables. We introduce REaLTabFormer (Realistic Relational and Tabular Transformer), a tabular and relational synthetic data generation model. It first creates a parent table using an autoregressive GPT-2 model, then generates the relational dataset conditioned on the parent table using a sequence-to-sequence (Seq2Seq) model. We implement target masking to prevent data copying and propose the $Q_{\delta}$ statistic and statistical bootstrapping to detect overfitting. Experiments using real-world datasets show that REaLTabFormer captures the relational structure better than a baseline model. REaLTabFormer also achieves state-of-the-art results on prediction tasks,"out-of-the-box", for large non-relational datasets without needing fine-tuning.

2023: Aivin Solatorio, Olivier Dupriez

https://arxiv.org/pdf/2302.02041v1.pdf

With the urgent demand for generalized deep models, many pre-trained big models are proposed, such as BERT, ViT, GPT, etc. Inspired by the success of these models in single domains (like computer vision and natural language processing), the multi-modal pre-trained big models have also drawn more and more attention in recent years. In this work, we give a comprehensive survey of these models and hope this paper could provide new insights and helps fresh researchers to track the most cutting-edge works. Specifically, we firstly introduce the background of multi-modal pre-training by reviewing the conventional deep learning, pre-training works in natural language process, computer vision, and speech.

2023: Xiao Wang, Guangyao Chen, Guangwu Qian, Pengcheng Gao, Xiaoyong Wei, Yaowei Wang, Yonghong Tian, Wen Gao

https://arxiv.org/pdf/2302.10035v1.pdf

Reward learning enables the application of reinforcement learning (RL) to tasks where reward is defined by human judgment, building a model of reward by asking humans questions. Most work on reward learning has used simulated environments, but complex information about values is often expressed in natural language, and we believe reward learning for language is a key to making RL practical and safe for real-world tasks. In this paper, we build on advances in generative pretraining of language models to apply reward learning to four natural language tasks: continuing text with positive sentiment or physically descriptive language, and summarization tasks on the TL;DR and CNN/Daily Mail datasets. For stylistic continuation we achieve good results with only 5,000 comparisons evaluated by humans. For summarization, models trained with 60,000 comparisons copy whole sentences from the input but skip irrelevant preamble; this leads to reasonable ROUGE scores and very good performance according to our human labelers, but may be exploiting the fact that labelers rely on simple heuristics.

2019: Daniel M. Ziegler, Nisan Stiennon, Jeff Wu, Tom B. Brown, Alec Radford, Dario Amodei, Paul Christiano, Geoffrey Irving

https://arxiv.org/pdf/1909.08593v2.pdf

In this work, we present a conceptually sim-ple and effective method to train a strong bilingual/multilingual multimodal representation model. Starting from the pre-trained multimodal representation model CLIP released by OpenAI, we altered its text encoder with a pre-trained multilingual text encoder XLMR, and aligned both languages and image representations by a two-stage training schema consisting of teacher learning and contrastive learning. We validate our method through evaluations of a wide range of tasks. We set new state-of-the-art performances on a bunch of tasks including ImageNet-CN, Flicker30k-CN, COCO-CN and XTD. Further, we obtain very close performances with CLIP on almost all tasks, suggesting that one can simply alter the text encoder in CLIP for extended capabilities such as multilingual understanding.

2022: Zhongzhi Chen, Guangyi Liu, Bo Zhang, Fulong Ye, Qinghong Yang, Ledell Yu Wu

Ranked #1 on Zero-Shot Transfer Image Classification on CN-ImageNet-Sketch

https://arxiv.org/pdf/2211.06679v2.pdf

Retrieval-augmented in-context learning has emerged as a powerful approach for addressing knowledge-intensive tasks using frozen language models (LM) and retrieval models (RM). Existing work has combined these in simple “retrieve-then-read” pipelines in which the RM retrieves passages that are inserted into the LM prompt. To begin to fully realize the potential of frozen LMs and RMs, we propose D EMONSTRATE – S EARCH –P REDICT (DSP), a framework that relies on passing natural language texts in sophisticated pipelines between an LM and an RM. DSP can express high-level programs that bootstrap pipeline-aware demonstrations, search for relevant passages, and generate grounded predictions, systematically breaking down problems into small transformations that the LM and RM can handle more reliably. We have written novel DSP programs for answering questions in open-domain, multi-hop, and conversational settings, establish-ing in early evaluations new state-of-the-art in-context learning results and delivering 37–120%, 8–39%, and 80–290% relative gains against the vanilla LM (GPT-3.5), a standard retrieve-then-read pipeline, and a contemporaneous self-ask pipeline, respectively. We release DSP at https: //github.com/stanfordnlp/dsp .

2022: O. Khattab, Keshav Santhanam, Xiang Lisa Li, D. Hall, Percy Liang, Christopher Potts, M. Zaharia

https://arxiv.org/pdf/2212.14024v2.pdf

Retrieval-augmented in-context learning has emerged as a powerful approach for addressing knowledge-intensive tasks using frozen language models (LM) and retrieval models (RM). Existing work has combined these in simple “retrieve-then-read” pipelines in which the RM retrieves passages that are inserted into the LM prompt. To begin to fully realize the potential of frozen LMs and RMs, we propose D EMONSTRATE – S EARCH –P REDICT (DSP), a framework that relies on passing natural language texts in sophisticated pipelines between an LM and an RM. DSP can express high-level programs that bootstrap pipeline-aware demonstrations, search for relevant passages, and generate grounded predictions, systematically breaking down problems into small transformations that the LM and RM can handle more reliably. We have written novel DSP programs for answering questions in open-domain, multi-hop, and conversational settings, establish-ing in early evaluations new state-of-the-art in-context learning results and delivering 37–120%, 8–39%, and 80–290% relative gains against the vanilla LM (GPT-3.5), a standard retrieve-then-read pipeline, and a contemporaneous self-ask pipeline, respectively. We release DSP at https: //github.com/stanfordnlp/dsp .

2022: O. Khattab, Keshav Santhanam, Xiang Lisa Li, D. Hall, Percy Liang, Christopher Potts, M. Zaharia

https://arxiv.org/pdf/2212.14024v2.pdf