The current trend of scaling language models involves increasing both parameter count and training dataset size. Extrapolating this trend suggests that training dataset size may soon be limited by the amount of text data available on the internet. Motivated by this limit, we investigate scaling language models in data-constrained regimes. Specifically, we run a large set of experiments varying the extent of data repetition and compute budget, ranging up to 900 billion training tokens and 9 billion parameter models. We find that with constrained data for a fixed compute budget, training with up to 4 epochs of repeated data yields negligible changes to loss compared to having unique data. However, with more repetition, the value of adding compute eventually decays to zero. We propose and empirically validate a scaling law for compute optimality that accounts for the decreasing value of repeated tokens and excess parameters. Finally, we experiment with approaches mitigating data scarcity, including augmenting the training dataset with code data or removing commonly used filters. Models and datasets from our 400 training runs are publicly available at //github.com/huggingface/datablations.
相關內容
Supply chain operations traditionally involve a variety of complex decision making problems. Over the last few decades, supply chains greatly benefited from advances in computation, which allowed the transition from manual processing to automation and cost-effective optimization. Nonetheless, business operators still need to spend substantial efforts in explaining and interpreting the optimization outcomes to stakeholders. Motivated by the recent advances in Large Language Models (LLMs), we study how this disruptive technology can help bridge the gap between supply chain automation and human comprehension and trust thereof. We design OptiGuide -- a framework that accepts as input queries in plain text, and outputs insights about the underlying optimization outcomes. Our framework does not forgo the state-of-the-art combinatorial optimization technology, but rather leverages it to quantitatively answer what-if scenarios (e.g., how would the cost change if we used supplier B instead of supplier A for a given demand?). Importantly, our design does not require sending proprietary data over to LLMs, which can be a privacy concern in some circumstances. We demonstrate the effectiveness of our framework on a real server placement scenario within Microsoft's cloud supply chain. Along the way, we develop a general evaluation benchmark, which can be used to evaluate the accuracy of the LLM output in other scenarios.
Training Artificial Intelligence (AI) models on three-dimensional image data presents unique challenges compared to the two-dimensional case: Firstly, the computational resources are significantly higher, and secondly, the availability of large pretraining datasets is often limited, impeding training success. In this study, we propose a simple approach of adapting 2D networks with an intermediate feature representation for processing 3D volumes. Our method involves sequentially applying these networks to slices of a 3D volume from all orientations. Subsequently, a feature reduction module combines the extracted slice features into a single representation, which is then used for classification. We evaluate our approach on medical classification benchmarks and a real-world clinical dataset, demonstrating comparable results to existing methods. Furthermore, by employing attention pooling as a feature reduction module we obtain weighted importance values for each slice during the forward pass. We show that slices deemed important by our approach allow the inspection of the basis of a model's prediction.
Large language models (LLMs) demonstrate remarkable ability to comprehend, reason, and generate following nature language instructions. However, the development of LLMs has been primarily focused on high-resource languages, such as English, thereby limiting their applicability and research in other languages. Consequently, we present PolyLM, a multilingual LLM trained on 640 billion (B) tokens, avaliable in two model sizes: 1.7B and 13B. To enhance its multilingual capabilities, we 1) integrate bilingual data into training data; and 2) adopt a curriculum learning strategy that increases the proportion of non-English data from 30% in the first stage to 60% in the final stage during pre-training. Further, we propose a multilingual self-instruct method which automatically generates 132.7K diverse multilingual instructions for model fine-tuning. To assess the model's performance, we collect several existing multilingual tasks, including multilingual understanding, question answering, generation, and translation. Extensive experiments show that PolyLM surpasses other open-source models such as LLaMA and BLOOM on multilingual tasks while maintaining comparable performance in English. Our models, alone with the instruction data and multilingual benchmark, are available at: \url{//modelscope.cn/models/damo/nlp_polylm_13b_text_generation}.
Denoising diffusion models have recently emerged as the predominant paradigm for generative modelling. Their extension to Riemannian manifolds has facilitated their application to an array of problems in the natural sciences. Yet, in many practical settings, such manifolds are defined by a set of constraints and are not covered by the existing (Riemannian) diffusion model methodology. Recent work has attempted to address this issue by employing novel noising processes based on logarithmic barrier methods or reflected Brownian motions. However, the associated samplers are computationally burdensome as the complexity of the constraints increases. In this paper, we introduce an alternative simple noising scheme based on Metropolis sampling that affords substantial gains in computational efficiency and empirical performance compared to the earlier samplers. Of independent interest, we prove that this new process corresponds to a valid discretisation of the reflected Brownian motion. We demonstrate the scalability and flexibility of our approach on a range of problem settings with convex and non-convex constraints, including applications from geospatial modelling, robotics and protein design.
Statistical shape models (SSM) have been well-established as an excellent tool for identifying variations in the morphology of anatomy across the underlying population. Shape models use consistent shape representation across all the samples in a given cohort, which helps to compare shapes and identify the variations that can detect pathologies and help in formulating treatment plans. In medical imaging, computing these shape representations from CT/MRI scans requires time-intensive preprocessing operations, including but not limited to anatomy segmentation annotations, registration, and texture denoising. Deep learning models have demonstrated exceptional capabilities in learning shape representations directly from volumetric images, giving rise to highly effective and efficient Image-to-SSM. Nevertheless, these models are data-hungry and due to the limited availability of medical data, deep learning models tend to overfit. Offline data augmentation techniques, that use kernel density estimation based (KDE) methods for generating shape-augmented samples, have successfully aided Image-to-SSM networks in achieving comparable accuracy to traditional SSM methods. However, these augmentation methods focus on shape augmentation, whereas deep learning models exhibit image-based texture bias results in sub-optimal models. This paper introduces a novel strategy for on-the-fly data augmentation for the Image-to-SSM framework by leveraging data-dependent noise generation or texture augmentation. The proposed framework is trained as an adversary to the Image-to-SSM network, augmenting diverse and challenging noisy samples. Our approach achieves improved accuracy by encouraging the model to focus on the underlying geometry rather than relying solely on pixel values.
We present this article as a small gesture in an attempt to counter what appears to be exponentially growing hype around Artificial Intelligence (AI) and its capabilities, and the distraction provided by the associated talk of science-fiction scenarios that might arise if AI should become sentient and super-intelligent. It may also help those outside of the field to become more informed about some of the limitations of AI technology. In the current context of popular discourse AI defaults to mean foundation and large language models (LLMs) such as those used to create ChatGPT. This in itself is a misrepresentation of the diversity, depth and volume of research, researchers, and technology that truly represents the field of AI. AI being a field of research that has existed in software artefacts since at least the 1950's. We set out to highlight a number of limitations of LLMs, and in so doing highlight that harms have already arisen and will continue to arise due to these limitations. Along the way we also highlight some of the associated risks for individuals and organisations in using this technology.
With the rise of powerful pre-trained vision-language models like CLIP, it becomes essential to investigate ways to adapt these models to downstream datasets. A recently proposed method named Context Optimization (CoOp) introduces the concept of prompt learning -- a recent trend in NLP -- to the vision domain for adapting pre-trained vision-language models. Specifically, CoOp turns context words in a prompt into a set of learnable vectors and, with only a few labeled images for learning, can achieve huge improvements over intensively-tuned manual prompts. In our study we identify a critical problem of CoOp: the learned context is not generalizable to wider unseen classes within the same dataset, suggesting that CoOp overfits base classes observed during training. To address the problem, we propose Conditional Context Optimization (CoCoOp), which extends CoOp by further learning a lightweight neural network to generate for each image an input-conditional token (vector). Compared to CoOp's static prompts, our dynamic prompts adapt to each instance and are thus less sensitive to class shift. Extensive experiments show that CoCoOp generalizes much better than CoOp to unseen classes, even showing promising transferability beyond a single dataset; and yields stronger domain generalization performance as well. Code is available at //github.com/KaiyangZhou/CoOp.
The dominating NLP paradigm of training a strong neural predictor to perform one task on a specific dataset has led to state-of-the-art performance in a variety of applications (eg. sentiment classification, span-prediction based question answering or machine translation). However, it builds upon the assumption that the data distribution is stationary, ie. that the data is sampled from a fixed distribution both at training and test time. This way of training is inconsistent with how we as humans are able to learn from and operate within a constantly changing stream of information. Moreover, it is ill-adapted to real-world use cases where the data distribution is expected to shift over the course of a model's lifetime. The first goal of this thesis is to characterize the different forms this shift can take in the context of natural language processing, and propose benchmarks and evaluation metrics to measure its effect on current deep learning architectures. We then proceed to take steps to mitigate the effect of distributional shift on NLP models. To this end, we develop methods based on parametric reformulations of the distributionally robust optimization framework. Empirically, we demonstrate that these approaches yield more robust models as demonstrated on a selection of realistic problems. In the third and final part of this thesis, we explore ways of efficiently adapting existing models to new domains or tasks. Our contribution to this topic takes inspiration from information geometry to derive a new gradient update rule which alleviate catastrophic forgetting issues during adaptation.
The recent GPT-3 model (Brown et al., 2020) achieves remarkable few-shot performance solely by leveraging a natural-language prompt and a few task demonstrations as input context. Inspired by their findings, we study few-shot learning in a more practical scenario, where we use smaller language models for which fine-tuning is computationally efficient. We present LM-BFF--better few-shot fine-tuning of language models--a suite of simple and complementary techniques for fine-tuning language models on a small number of annotated examples. Our approach includes (1) prompt-based fine-tuning together with a novel pipeline for automating prompt generation; and (2) a refined strategy for dynamically and selectively incorporating demonstrations into each context. Finally, we present a systematic evaluation for analyzing few-shot performance on a range of NLP tasks, including classification and regression. Our experiments demonstrate that our methods combine to dramatically outperform standard fine-tuning procedures in this low resource setting, achieving up to 30% absolute improvement, and 11% on average across all tasks. Our approach makes minimal assumptions on task resources and domain expertise, and hence constitutes a strong task-agnostic method for few-shot learning.
Language model based pre-trained models such as BERT have provided significant gains across different NLP tasks. In this paper, we study different types of pre-trained transformer based models such as auto-regressive models (GPT-2), auto-encoder models (BERT), and seq2seq models (BART) for conditional data augmentation. We show that prepending the class labels to text sequences provides a simple yet effective way to condition the pre-trained models for data augmentation. On three classification benchmarks, pre-trained Seq2Seq model outperforms other models. Further, we explore how different pre-trained model based data augmentation differs in-terms of data diversity, and how well such methods preserve the class-label information.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191