Training state-of-the-art neural networks requires a high cost in terms of compute and time. Model scale is recognized to be a critical factor to achieve and improve the state-of-the-art. Increasing the scale of a neural network normally requires restarting from scratch by randomly initializing all the parameters of the model, as this implies a change of architecture's parameters that does not allow for a straightforward transfer of knowledge from smaller size models. In this work, we propose six composable transformations to incrementally increase the size of transformer-based neural networks while preserving functionality, allowing to expand the capacity of the model as needed. We provide proof of exact function preservation under minimal initialization constraints for each transformation. The proposed methods may enable efficient training pipelines for larger and more powerful models by progressively expanding the architecture throughout training.

A standard ML model is commonly generated by a single method that specifies aspects such as architecture, initialization, training data and hyperparameters configuration. The presented work introduces a novel methodology allowing to define multiple methods as distinct agents. Agents can collaborate and compete to generate and improve ML models for a given tasks. The proposed methodology is demonstrated with the generation and extension of a dynamic modular multitask ML system solving more than one hundred image classification tasks. Diverse agents can compete to produce the best performing model for a task by reusing the modules introduced to the system by competing agents. The presented work focuses on the study of agents capable of: 1) reusing the modules generated by concurrent agents, 2) activating in parallel multiple modules in a frozen state by connecting them with trainable modules, 3) condition the activation mixture on each data sample by using a trainable router module. We demonstrate that this simple per-sample parallel routing method can boost the quality of the combined solutions by training a fraction of the activated parameters.

The traditional ML development methodology does not enable a large number of contributors, each with distinct objectives, to work collectively on the creation and extension of a shared intelligent system. Enabling such a collaborative methodology can accelerate the rate of innovation, increase ML technologies accessibility and enable the emergence of novel capabilities. We believe that this novel methodology for ML development can be demonstrated through a modularized representation of ML models and the definition of novel abstractions allowing to implement and execute diverse methods for the asynchronous use and extension of modular intelligent systems. We present a multiagent framework for the collaborative and asynchronous extension of dynamic large-scale multitask systems.

Multitask learning assumes that models capable of learning from multiple tasks can achieve better quality and efficiency via knowledge transfer, a key feature of human learning. Though, state of the art ML models rely on high customization for each task and leverage size and data scale rather than scaling the number of tasks. Also, continual learning, that adds the temporal aspect to multitask, is often focused to the study of common pitfalls such as catastrophic forgetting instead of being studied at a large scale as a critical component to build the next generation artificial intelligence.We propose an evolutionary method capable of generating large scale multitask models that support the dynamic addition of new tasks. The generated multitask models are sparsely activated and integrates a task-based routing that guarantees bounded compute cost and fewer added parameters per task as the model expands.The proposed method relies on a knowledge compartmentalization technique to achieve immunity against catastrophic forgetting and other common pitfalls such as gradient interference and negative transfer. We demonstrate empirically that the proposed method can jointly solve and achieve competitive results on 69public image classification tasks, for example improving the state of the art on a competitive benchmark such as cifar10 by achieving a 15% relative error reduction compared to the best model trained on public data.

The core challenge with continual learning is catastrophic forgetting, the phenomenon that when neural networks are trained on a sequence of tasks they rapidly forget previously learned tasks. It has been observed that catastrophic forgetting is most severe when tasks are dissimilar to each other. We propose the use of sparse routing networks for continual learning. For each input, these network architectures activate a different path through a network of experts. Routing networks have been shown to learn to route similar tasks to overlapping sets of experts and dissimilar tasks to disjoint sets of experts. In the continual learning context this behaviour is desirable as it minimizes interference between dissimilar tasks while allowing positive transfer between related tasks. In practice, we find it is necessary to develop a new training method for routing networks, which we call co-training which avoids poorly initialized experts when new tasks are presented. When combined with a small episodic memory replay buffer, sparse routing networks with co-training outperform densely connected networks on the MNIST-Permutations and MNIST-Rotations benchmarks.

A BSTRACT This paper proposes a novel per-task routing method for multi-task applications. Multi-task neural networks can learn to transfer knowledge across different tasks by using parameter sharing. However, sharing parameters between unrelated tasks can hurt performance. To address this issue, we advocate the use of routing networks to learn flexible parameter sharing, where each group of parameters is shared with a different subset of tasks in order to better leverage tasks relatedness. At the same time, it is known that routing networks are notoriously hard to train. We propose the Gumbel-Matrix routing: a novel multi-task routing method, designed to learn fine-grained patterns of parameter sharing. The routing is learned jointly with the model parameters by standard back-propagation thanks to the Gumbel-Softmax trick. When applied to the Omniglot benchmark, the proposed method reduces the state-of-the-art error rate by 17% . 1 I NTRODUCTION Multi-task learning (Caruana, 1998; 1993) based on neural networks has attracted lots of research interest in the past years and has been successfully applied to several application domains, such as recommender systems (Bansal et al., 2016) and real-time object detection (Girshick, 2015). For instance, a movie recommendation system may optimize not only the likelihood of the user clicking on a suggested movie, but also the likelihood that the user is going to watch it. The most common architecture used in practice for multi-task learning is the so-called shared bottom, where the tasks share parameters in the early layers of the model, which are followed by task-specific heads. However, as our experiments on synthetic data show, when the tasks are unrelated, parameter sharing may actually hurt individual tasks performance. Therefore, resorting to flexible parameter sharing becomes very important.

Recent advances in Neural Architecture Search (NAS) have produced state-of-the-art architectures on several tasks. NAS shifts the efforts of human experts from developing novel architectures directly to designing architecture search spaces and methods to explore them efficiently. The search space definition captures prior knowledge about the properties of the architectures and it is crucial for the complexity and the performance of the search algorithm. However, different search space definitions require restarting the learning process from scratch. We propose a novel agent based on the Transformer that supports joint training and efficient transfer of prior knowledge between multiple search spaces and tasks.

Neural architecture search has been shown to hold great promise towards the automation of deep learning. However in spite of its potential, neural architecture search remains quite costly. To this point, we propose a novel gradient-based framework for efficient architecture search by sharing information across several tasks. We start by training many model architectures on several related (training) tasks. When a new unseen task is presented, the framework performs architecture inference in order to quickly identify a good candidate architecture, before any model is trained on the new task. At the core of our framework lies a deep value network that can predict the performance of input architectures on a task by utilizing task meta-features and the previous model training experiments performed on related tasks. We adopt a continuous parametrization of the model architecture which allows for efficient gradient-based optimization. Given a new task, an effective architecture is quickly identified by maximizing the estimated performance with respect to the model architecture parameters with simple gradient ascent. It is key to point out that our goal is to achieve reasonable performance at the lowest cost. We provide experimental results showing the effectiveness of the framework despite its high computational efficiency.

Fine-tuning large pre-trained models is an effective transfer mechanism in NLP. However, in the presence of many downstream tasks, fine-tuning is parameter inefficient: an entire new model is required for every task. As an alternative, we propose transfer with adapter modules. Adapter modules yield a compact and extensible model; they add only a few trainable parameters per task, and new tasks can be added without revisiting previous ones. The parameters of the original network remain fixed, yielding a high degree of parameter sharing. To demonstrate adapter's effectiveness, we transfer the recently proposed BERT Transformer model to 26 diverse text classification tasks, including the GLUE benchmark. Adapters attain near state-of-the-art performance, whilst adding only a few parameters per task. On GLUE, we attain within 0.4% of the performance of full fine-tuning, adding only 3.6% parameters per task. By contrast, fine-tuning trains 100% of the parameters per task.

Neural Architecture Search has recently shown potential to automate the design of Neural Networks. The use of Neural Network agents trained with Reinforcement Learning can offer the possibility to learn complex architectural patterns, as well as the ability to explore a vast and compositional search space. On the other hand, evolutionary algorithms offer the sample efficiency needed for such a resource intensive application. We propose a class of Evolutionary-Neural hybrid agents (Evo-NAS), that retain the qualities of the two approaches. We show that the Evo-NAS agent outperforms both Neural and Evolutionary agents when applied to architecture search for a suite of text classification and image classification benchmarks. On a high-complexity architecture search space for image classification, the Evo-NAS agent surpasses the performance of commonly used agents with only 1/3 of the trials.