Large Language Models have demonstrated remarkable few-shot performance, but the performance can be sensitive to the selection of few-shot instances. We propose PATRON, a new method that uses prompt-based uncertainty estimation for data selection for pre-trained language model fine-tuning under cold-start scenarios, i.e., no initial labeled data are available. In PATRON, we design (1) a prompt-based uncertainty propagation approach to estimate the importance of data points and (2) a partition-then-rewrite (PTR) strategy to promote sample diversity when querying for annotations. Experiments on six text classification datasets show that PATRON outperforms the strongest cold-start data selection baselines by up to 6.9%. Besides, with 128 labels only, PATRON achieves 91.0% and 92.1% of the fully supervised performance based on vanilla fine-tuning and prompt-based learning respectively. Our implementation of PATRON is available at \url{//github.com/yueyu1030/Patron}.
相關內容
Gaussian process (GP) is a Bayesian model which provides several advantages for regression tasks in machine learning such as reliable quantitation of uncertainty and improved interpretability. Their adoption has been precluded by their excessive computational cost and by the difficulty in adapting them for analyzing sequences (e.g. amino acid and nucleotide sequences) and graphs (e.g. ones representing small molecules). In this study, we develop efficient and scalable approaches for fitting GP models as well as fast convolution kernels which scale linearly with graph or sequence size. We implement these improvements by building an open-source Python library called xGPR. We compare the performance of xGPR with the reported performance of various deep learning models on 20 benchmarks, including small molecule, protein sequence and tabular data. We show that xGRP achieves highly competitive performance with much shorter training time. Furthermore, we also develop new kernels for sequence and graph data and show that xGPR generally outperforms convolutional neural networks on predicting key properties of proteins and small molecules. Importantly, xGPR provides uncertainty information not available from typical deep learning models. Additionally, xGPR provides a representation of the input data that can be used for clustering and data visualization. These results demonstrate that xGPR provides a powerful and generic tool that can be broadly useful in protein engineering and drug discovery.
In this paper we explore few-shot imitation learning for control problems, which involves learning to imitate a target policy by accessing a limited set of offline rollouts. This setting has been relatively under-explored despite its relevance to robotics and control applications. State-of-the-art methods developed to tackle few-shot imitation rely on meta-learning, which is expensive to train as it requires access to a distribution over tasks (rollouts from many target policies and variations of the base environment). Given this limitation we investigate an alternative approach, fine-tuning, a family of methods that pretrain on a single dataset and then fine-tune on unseen domain-specific data. Recent work has shown that fine-tuners outperform meta-learners in few-shot image classification tasks, especially when the data is out-of-domain. Here we evaluate to what extent this is true for control problems, proposing a simple yet effective baseline which relies on two stages: (i) training a base policy online via reinforcement learning (e.g. Soft Actor-Critic) on a single base environment, (ii) fine-tuning the base policy via behavioral cloning on a few offline rollouts of the target policy. Despite its simplicity this baseline is competitive with meta-learning methods on a variety of conditions and is able to imitate target policies trained on unseen variations of the original environment. Importantly, the proposed approach is practical and easy to implement, as it does not need any complex meta-training protocol. As a further contribution, we release an open source dataset called iMuJoCo (iMitation MuJoCo) consisting of 154 variants of popular OpenAI-Gym MuJoCo environments with associated pretrained target policies and rollouts, which can be used by the community to study few-shot imitation learning and offline reinforcement learning.
We present an artificial intelligence (AI) method for automatically computing the melting point based on coexistence simulations in the NPT ensemble. Given the interatomic interaction model, the method makes decisions regarding the number of atoms and temperature at which to conduct simulations, and based on the collected data predicts the melting point along with the uncertainty, which can be systematically improved with more data. We demonstrate how incorporating physical models of the solid-liquid coexistence evolution enhances the AI method's accuracy and enables optimal decision-making to effectively reduce predictive uncertainty. To validate our approach, we compare our results with approximately 20 melting point calculations from the literature. Remarkably, we observe significant deviations in about one-third of the cases, underscoring the need for accurate and reliable AI-based algorithms for materials property calculations.
Message-passing Graph Neural Networks (GNNs), which collect information from adjacent nodes achieve dismal performance on heterophilic graphs. Various schemes have been proposed to solve this problem, and propagating signed information on heterophilic edges has gained great attention. Recently, some works provided theoretical analysis that signed propagation always leads to performance improvement under a binary class scenario. However, we notice that prior analyses do not align well with multi-class benchmark datasets. This paper provides a new understanding of signed propagation for multi-class scenarios and points out two drawbacks in terms of message-passing and parameter update: (1) Message-passing: if two nodes belong to different classes but have a high similarity, signed propagation can decrease the separability. (2) Parameter update: the prediction uncertainty (e.g., conflict evidence) of signed neighbors increases during training, which can impede the stability of the algorithm. Based on the observation, we introduce two novel strategies for improving signed propagation under multi-class graphs. The proposed scheme combines calibration to secure robustness while reducing uncertainty. We show the efficacy of our theorem through extensive experiments on six benchmark graph datasets.
In this work, we propose a built-in Deep Learning Physics Optimization (DLPO) framework to set up a shape optimization study of the Duisburg Test Case (DTC) container vessel. We present two different applications: (1) sensitivity analysis to detect the most promising generic basis hull shapes, and (2) multi-objective optimization to quantify the trade-off between optimal hull forms. DLPO framework allows for the evaluation of design iterations automatically in an end-to-end manner. We achieved these results by coupling Extrality's Deep Learning Physics (DLP) model to a CAD engine and an optimizer. Our proposed DLP model is trained on full 3D volume data coming from RANS simulations, and it can provide accurate and high-quality 3D flow predictions in real-time, which makes it a good evaluator to perform optimization of new container vessel designs w.r.t the hydrodynamic efficiency. In particular, it is able to recover the forces acting on the vessel by integration on the hull surface with a mean relative error of 3.84\% \pm 2.179\% on the total resistance. Each iteration takes only 20 seconds, thus leading to a drastic saving of time and engineering efforts, while delivering valuable insight into the performance of the vessel, including RANS-like detailed flow information. We conclude that DLPO framework is a promising tool to accelerate the ship design process and lead to more efficient ships with better hydrodynamic performance.
Despite the significant interest and progress in reinforcement learning (RL) problems with adversarial corruption, current works are either confined to the linear setting or lead to an undesired $\tilde{O}(\sqrt{T}\zeta)$ regret bound, where $T$ is the number of rounds and $\zeta$ is the total amount of corruption. In this paper, we consider the contextual bandit with general function approximation and propose a computationally efficient algorithm to achieve a regret of $\tilde{O}(\sqrt{T}+\zeta)$. The proposed algorithm relies on the recently developed uncertainty-weighted least-squares regression from linear contextual bandit and a new weighted estimator of uncertainty for the general function class. In contrast to the existing analysis that heavily relies on the linear structure, we develop a novel technique to control the sum of weighted uncertainty, thus establishing the final regret bounds. We then generalize our algorithm to the episodic MDP setting and first achieve an additive dependence on the corruption level $\zeta$ in the scenario of general function approximation. Notably, our algorithms achieve regret bounds either nearly match the performance lower bound or improve the existing methods for all the corruption levels and in both known and unknown $\zeta$ cases.
Unsupervised domain adaptation has recently emerged as an effective paradigm for generalizing deep neural networks to new target domains. However, there is still enormous potential to be tapped to reach the fully supervised performance. In this paper, we present a novel active learning strategy to assist knowledge transfer in the target domain, dubbed active domain adaptation. We start from an observation that energy-based models exhibit free energy biases when training (source) and test (target) data come from different distributions. Inspired by this inherent mechanism, we empirically reveal that a simple yet efficient energy-based sampling strategy sheds light on selecting the most valuable target samples than existing approaches requiring particular architectures or computation of the distances. Our algorithm, Energy-based Active Domain Adaptation (EADA), queries groups of targe data that incorporate both domain characteristic and instance uncertainty into every selection round. Meanwhile, by aligning the free energy of target data compact around the source domain via a regularization term, domain gap can be implicitly diminished. Through extensive experiments, we show that EADA surpasses state-of-the-art methods on well-known challenging benchmarks with substantial improvements, making it a useful option in the open world. Code is available at //github.com/BIT-DA/EADA.
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.
Meta-learning extracts the common knowledge acquired from learning different tasks and uses it for unseen tasks. It demonstrates a clear advantage on tasks that have insufficient training data, e.g., few-shot learning. In most meta-learning methods, tasks are implicitly related via the shared model or optimizer. In this paper, we show that a meta-learner that explicitly relates tasks on a graph describing the relations of their output dimensions (e.g., classes) can significantly improve the performance of few-shot learning. This type of graph is usually free or cheap to obtain but has rarely been explored in previous works. We study the prototype based few-shot classification, in which a prototype is generated for each class, such that the nearest neighbor search between the prototypes produces an accurate classification. We introduce "Gated Propagation Network (GPN)", which learns to propagate messages between prototypes of different classes on the graph, so that learning the prototype of each class benefits from the data of other related classes. In GPN, an attention mechanism is used for the aggregation of messages from neighboring classes, and a gate is deployed to choose between the aggregated messages and the message from the class itself. GPN is trained on a sequence of tasks from many-shot to few-shot generated by subgraph sampling. During training, it is able to reuse and update previously achieved prototypes from the memory in a life-long learning cycle. In experiments, we change the training-test discrepancy and test task generation settings for thorough evaluations. GPN outperforms recent meta-learning methods on two benchmark datasets in all studied cases.
To address the sparsity and cold start problem of collaborative filtering, researchers usually make use of side information, such as social networks or item attributes, to improve recommendation performance. This paper considers the knowledge graph as the source of side information. To address the limitations of existing embedding-based and path-based methods for knowledge-graph-aware recommendation, we propose Ripple Network, an end-to-end framework that naturally incorporates the knowledge graph into recommender systems. Similar to actual ripples propagating on the surface of water, Ripple Network stimulates the propagation of user preferences over the set of knowledge entities by automatically and iteratively extending a user's potential interests along links in the knowledge graph. The multiple "ripples" activated by a user's historically clicked items are thus superposed to form the preference distribution of the user with respect to a candidate item, which could be used for predicting the final clicking probability. Through extensive experiments on real-world datasets, we demonstrate that Ripple Network achieves substantial gains in a variety of scenarios, including movie, book and news recommendation, over several state-of-the-art baselines.
小貼士
登錄享
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191