In settings where users are both time-pressured and need high accuracy, such as doctors working in Emergency Rooms, we want to provide AI assistance that both increases accuracy and reduces time. However, different types of AI assistance have different benefits: some reduce time taken while increasing overreliance on AI, while others do the opposite. We therefore want to adapt what AI assistance we show depending on various properties (of the question and of the user) in order to best tradeoff our two objectives. We introduce a study where users have to prescribe medicines to aliens, and use it to explore the potential for adapting AI assistance. We find evidence that it is beneficial to adapt our AI assistance depending on the question, leading to good tradeoffs between time taken and accuracy. Future work would consider machine-learning algorithms (such as reinforcement learning) to automatically adapt quickly.
相關內容
In this paper we revisit the efficacy of knowledge distillation as a function matching and metric learning problem. In doing so we verify three important design decisions, namely the normalisation, soft maximum function, and projection layers as key ingredients. We theoretically show that the projector implicitly encodes information on past examples, enabling relational gradients for the student. We then show that the normalisation of representations is tightly coupled with the training dynamics of this projector, which can have a large impact on the students performance. Finally, we show that a simple soft maximum function can be used to address any significant capacity gap problems. Experimental results on various benchmark datasets demonstrate that using these insights can lead to superior or comparable performance to state-of-the-art knowledge distillation techniques, despite being much more computationally efficient. In particular, we obtain these results across image classification (CIFAR100 and ImageNet), object detection (COCO2017), and on more difficult distillation objectives, such as training data efficient transformers, whereby we attain a 77.2% top-1 accuracy with DeiT-Ti on ImageNet.
We present our proposed solution to the BabyLM challenge [arXiv:2301.11796], whose goal was to improve the sample efficiency of language models. We trained an ensemble consisting of a GPT-2 and small LLaMA models on the developmentally-plausible, 10M-word BabyLM dataset, then distilled it into a small, 58M-parameter LLaMA model, which exceeds in performance both of its teachers as well as a similar model trained without distillation. This suggests that distillation can not only retain the full performance of the teacher model when the latter is trained on a sufficiently small dataset; it can exceed it, and lead to significantly better performance than direct training.
In order to fully harness the potential of dielectric elastomer actu-ators (DEAs) in soft robots, advanced control methods are need-ed. An important groundwork for this is the development of a control-oriented model that can adequately describe the underly-ing dynamics of a DEA. A common feature of existing models is that always custom-made DEAs were investigated. This makes the modelling process easier, as all specifications and the struc-ture of the actuator are well known. In the case of a commercial actuator, however, only the information from the manufacturer is available and must be checked or completed during the modelling process. The aim of this paper is to explore how a commercial stacked silicone-based DEA can be modelled and how complex the model should be to properly replicate the features of the actu-ator. The static description has demonstrated the suitability of Hooke's law. In the case of dynamic description, it is shown that no viscoelastic model is needed for control-oriented modelling. However, if all features of the DEA are considered, the general-ized Kelvin-Maxwell model with three Maxwell elements shows good results, stability and computational efficiency.
In this paper, we develop a unified regression approach to model unconditional quantiles, M-quantiles and expectiles of multivariate dependent variables exploiting the multidimensional Huber's function. To assess the impact of changes in the covariates across the entire unconditional distribution of the responses, we extend the work of Firpo et al. (2009) by running a mean regression of the recentered influence function on the explanatory variables. We discuss the estimation procedure and establish the asymptotic properties of the derived estimators. A data-driven procedure is also presented to select the tuning constant of the Huber's function. The validity of the proposed methodology is explored with simulation studies and through an application using the Survey of Household Income and Wealth 2016 conducted by the Bank of Italy.
This paper presents DeepTSF, a comprehensive machine learning operations (MLOps) framework aiming to innovate time series forecasting through workflow automation and codeless modeling. DeepTSF automates key aspects of the ML lifecycle, making it an ideal tool for data scientists and MLops engineers engaged in machine learning (ML) and deep learning (DL)-based forecasting. DeepTSF empowers users with a robust and user-friendly solution, while it is designed to seamlessly integrate with existing data analysis workflows, providing enhanced productivity and compatibility. The framework offers a front-end user interface (UI) suitable for data scientists, as well as other higher-level stakeholders, enabling comprehensive understanding through insightful visualizations and evaluation metrics. DeepTSF also prioritizes security through identity management and access authorization mechanisms. The application of DeepTSF in real-life use cases of the I-NERGY project has already proven DeepTSF's efficacy in DL-based load forecasting, showcasing its significant added value in the electrical power and energy systems domain.
Buy It Again (BIA) recommendations are crucial to retailers to help improve user experience and site engagement by suggesting items that customers are likely to buy again based on their own repeat purchasing patterns. Most existing BIA studies analyze guests personalized behavior at item granularity. A category-based model may be more appropriate in such scenarios. We propose a recommendation system called a hierarchical PCIC model that consists of a personalized category model (PC model) and a personalized item model within categories (IC model). PC model generates a personalized list of categories that customers are likely to purchase again. IC model ranks items within categories that guests are likely to consume within a category. The hierarchical PCIC model captures the general consumption rate of products using survival models. Trends in consumption are captured using time series models. Features derived from these models are used in training a category-grained neural network. We compare PCIC to twelve existing baselines on four standard open datasets. PCIC improves NDCG up to 16 percent while improving recall by around 2 percent. We were able to scale and train (over 8 hours) PCIC on a large dataset of 100M guests and 3M items where repeat categories of a guest out number repeat items. PCIC was deployed and AB tested on the site of a major retailer, leading to significant gains in guest engagement.
In large-scale systems there are fundamental challenges when centralised techniques are used for task allocation. The number of interactions is limited by resource constraints such as on computation, storage, and network communication. We can increase scalability by implementing the system as a distributed task-allocation system, sharing tasks across many agents. However, this also increases the resource cost of communications and synchronisation, and is difficult to scale. In this paper we present four algorithms to solve these problems. The combination of these algorithms enable each agent to improve their task allocation strategy through reinforcement learning, while changing how much they explore the system in response to how optimal they believe their current strategy is, given their past experience. We focus on distributed agent systems where the agents' behaviours are constrained by resource usage limits, limiting agents to local rather than system-wide knowledge. We evaluate these algorithms in a simulated environment where agents are given a task composed of multiple subtasks that must be allocated to other agents with differing capabilities, to then carry out those tasks. We also simulate real-life system effects such as networking instability. Our solution is shown to solve the task allocation problem to 6.7% of the theoretical optimal within the system configurations considered. It provides 5x better performance recovery over no-knowledge retention approaches when system connectivity is impacted, and is tested against systems up to 100 agents with less than a 9% impact on the algorithms' performance.
This paper does not describe a working system. Instead, it presents a single idea about representation which allows advances made by several different groups to be combined into an imaginary system called GLOM. The advances include transformers, neural fields, contrastive representation learning, distillation and capsules. GLOM answers the question: How can a neural network with a fixed architecture parse an image into a part-whole hierarchy which has a different structure for each image? The idea is simply to use islands of identical vectors to represent the nodes in the parse tree. If GLOM can be made to work, it should significantly improve the interpretability of the representations produced by transformer-like systems when applied to vision or language
Recent work pre-training Transformers with self-supervised objectives on large text corpora has shown great success when fine-tuned on downstream NLP tasks including text summarization. However, pre-training objectives tailored for abstractive text summarization have not been explored. Furthermore there is a lack of systematic evaluation across diverse domains. In this work, we propose pre-training large Transformer-based encoder-decoder models on massive text corpora with a new self-supervised objective. In PEGASUS, important sentences are removed/masked from an input document and are generated together as one output sequence from the remaining sentences, similar to an extractive summary. We evaluated our best PEGASUS model on 12 downstream summarization tasks spanning news, science, stories, instructions, emails, patents, and legislative bills. Experiments demonstrate it achieves state-of-the-art performance on all 12 downstream datasets measured by ROUGE scores. Our model also shows surprising performance on low-resource summarization, surpassing previous state-of-the-art results on 6 datasets with only 1000 examples. Finally we validated our results using human evaluation and show that our model summaries achieve human performance on multiple datasets.
Nowadays, the Convolutional Neural Networks (CNNs) have achieved impressive performance on many computer vision related tasks, such as object detection, image recognition, image retrieval, etc. These achievements benefit from the CNNs outstanding capability to learn the input features with deep layers of neuron structures and iterative training process. However, these learned features are hard to identify and interpret from a human vision perspective, causing a lack of understanding of the CNNs internal working mechanism. To improve the CNN interpretability, the CNN visualization is well utilized as a qualitative analysis method, which translates the internal features into visually perceptible patterns. And many CNN visualization works have been proposed in the literature to interpret the CNN in perspectives of network structure, operation, and semantic concept. In this paper, we expect to provide a comprehensive survey of several representative CNN visualization methods, including Activation Maximization, Network Inversion, Deconvolutional Neural Networks (DeconvNet), and Network Dissection based visualization. These methods are presented in terms of motivations, algorithms, and experiment results. Based on these visualization methods, we also discuss their practical applications to demonstrate the significance of the CNN interpretability in areas of network design, optimization, security enhancement, etc.
注冊地址： 北京市海淀區羊坊店路18號2幢3層301-191