Purpose: We propose a novel method for continual learning based on the increasing depth of neural networks. This work explores whether extending neural network depth may be beneficial in a life-long learning setting. Methods: We propose a novel approach based on adding new layers on top of existing ones to enable the forward transfer of knowledge and adapting previously learned representations. We employ a method of determining the most similar tasks for selecting the best location in our network to add new nodes with trainable parameters. This approach allows for creating a tree-like model, where each node is a set of neural network parameters dedicated to a specific task. The Progressive Neural Network concept inspires the proposed method. Therefore, it benefits from dynamic changes in network structure. However, Progressive Neural Network allocates a lot of memory for the whole network structure during the learning process. The proposed method alleviates this by adding only part of a network for a new task and utilizing a subset of previously trained weights. At the same time, we may retain the benefit of PNN, such as no forgetting guaranteed by design, without needing a memory buffer. Results: Experiments on Split CIFAR and Split Tiny ImageNet show that the proposed algorithm is on par with other continual learning methods. In a more challenging setup with a single computer vision dataset as a separate task, our method outperforms Experience Replay. Conclusion: It is compatible with commonly used computer vision architectures and does not require a custom network structure. As an adaptation to changing data distribution is made by expanding the architecture, there is no need to utilize a rehearsal buffer. For this reason, our method could be used for sensitive applications where data privacy must be considered.

We introduce a physiological model-based agent as proof-of-principle that it is possible to define a flexible self-preserving system that does not use a reward signal or reward-maximization as an objective. We achieve this by introducing the Self-Preserving Agent (SPA) with a physiological structure where the system can get trapped in an absorbing state if the agent does not solve and execute goal-directed polices. Our agent is defined using new class of Bellman equations called Operator Bellman Equations (OBEs), for encoding jointly non-stationary non-Markovian tasks formalized as a Temporal Goal Markov Decision Process (TGMDP). OBEs produce optimal goal-conditioned spatiotemporal transition operators that map an initial state-time to the final state-times of a policy used to complete a goal, and can also be used to forecast future states in multiple dynamic physiological state-spaces. SPA is equipped with an intrinsic motivation function called the valence function, which quantifies the changes in empowerment (the channel capacity of a transition operator) after following a policy. Because empowerment is a function of a transition operator, there is a natural synergism between empowerment and OBEs: the OBEs create hierarchical transition operators, and the valence function can evaluate hierarchical empowerment change defined on these operators. The valence function can then be used for goal selection, wherein the agent chooses a policy sequence that realizes goal states which produce maximum empowerment gain. In doing so, the agent will seek freedom and avoid internal death-states that undermine its ability to control both external and internal states in the future, thereby exhibiting the capacity of predictive and anticipatory self-preservation. We also compare SPA to Multi-objective RL, and discuss its capacity for symbolic reasoning and life-long learning.

As companies use more and more data to improve how AI recognizes images, learns languages and carries out other complex tasks, a paper publishing in Science this week shows a way that computer chips could dynamically rewire themselves to take in new data like the brain does, helping AI to keep learning over time. "The brains of living beings can continuously learn throughout their lifespan. We have now created an artificial platform for machines to learn throughout their lifespan," said Shriram Ramanathan, a professor in Purdue University's School of Materials Engineering who specializes in discovering how materials could mimic the brain to improve computing. Unlike the brain, which constantly forms new connections between neurons to enable learning, the circuits on a computer chip don't change. A circuit that a machine has been using for years isn't any different than the circuit that was originally built for the machine in a factory.

When the human brain learns something new, it adapts. But when artificial intelligence learns something new, it tends to forget information it already learned. As companies use more and more data to improve how AI recognizes images, learns languages and carries out other complex tasks, a paper publishing in Science this week shows a way that computer chips could dynamically rewire themselves to take in new data like the brain does, helping AI to keep learning over time. "The brains of living beings can continuously learn throughout their lifespan. We have now created an artificial platform for machines to learn throughout their lifespan," said Shriram Ramanathan, a professor in Purdue University's School of Materials Engineering who specializes in discovering how materials could mimic the brain to improve computing.

Geneva: Are robots taking away human jobs? Where will the jobs of the future come from? Are our skills becoming increasingly outdated? Is the gig economy creating precarious jobs? The rapidly transforming world of work has made headlines of these social anxieties.

Semantic vectors are learned from data to express semantic relationships between elements of information, for the purpose of solving and informing downstream tasks. Other models exist that learn to map and classify supervised data. However, the two worlds of learning rarely interact to inform one another dynamically, whether across types of data or levels of semantics, in order to form a unified model. We explore the research problem of learning these vectors and propose a framework for learning the semantics of knowledge incrementally and online, across multiple mediums of data, via binary vectors. We discuss the aspects of this framework to spur future research on this approach and problem.

Learners’ skills decay during gaps in instruction, since they lack the structure and motivation to continue studying. To meet this challenge, the PAL3 system was designed to accompany a learner throughout their career and mentor them to build and maintain skills through: 1) the use of an embodied pedagogical agent (Pal), 2) a persistent learning record that drives a student model which estimates forgetting, 3) an adaptive recommendation engine linking to both intelligent tutors and traditional learning resources, and 4) game-like mechanisms to promote engagement (e.g., leaderboards, effort-based point rewards, unlocking customizations). The design process for PAL3 is discussed, from the perspective of insights and revisions based on a series of formative feedback and evaluation sessions.