arXiv subjects

Jürgen Schmidhuber

Publications and source records attributed to Jürgen Schmidhuber.

At least 91 records · Page 5Linked to original sources

arXiv2023

Accelerating Neural Self-Improvement via Bootstrapping

Few-shot learning with sequence-processing neural networks (NNs) has recently attracted a new wave of attention in the context of large language models. In the standard N-way K-shot learning setting, an NN is explicitly optimised to learn to classify unlabelled inputs by observing a sequence of NK labelled examples. This pressures the NN to learn a learning algorithm that achieves optimal performance, given the limited number of training examples. Here we study an auxiliary loss that encourages further acceleration of few-shot learning, by applying recently proposed bootstrapped meta-learning to NN few-shot learners: we optimise the K-shot learner to match its own performance achievable by observing more than NK examples, using only NK examples. Promising results are obtained on the standard Mini-ImageNet dataset. Our code is public.

cs.LG↗

arXiv2023

Guiding Online Reinforcement Learning with Action-Free Offline Pretraining

Offline RL methods have been shown to reduce the need for environment interaction by training agents using offline collected episodes. However, these methods typically require action information to be logged during data collection, which can be difficult or even impossible in some practical cases. In this paper, we investigate the potential of using action-free offline datasets to improve online reinforcement learning, name this problem Reinforcement Learning with Action-Free Offline Pretraining (AFP-RL). We introduce Action-Free Guide (AF-Guide), a method that guides online training by extracting knowledge from action-free offline datasets. AF-Guide consists of an Action-Free Decision Transformer (AFDT) implementing a variant of Upside-Down Reinforcement Learning. It learns to plan the next states from the offline dataset, and a Guided Soft Actor-Critic (Guided SAC) that learns online with guidance from AFDT. Experimental results show that AF-Guide can improve sample efficiency and performance in online training thanks to the knowledge from the action-free offline dataset. Code is available at https://github.com/Vision-CAIR/AF-Guide.

cs.LG↗

arXiv2023

Is it enough to optimize CNN architectures on ImageNet?

Classification performance based on ImageNet is the de-facto standard metric for CNN development. In this work we challenge the notion that CNN architecture design solely based on ImageNet leads to generally effective convolutional neural network (CNN) architectures that perform well on a diverse set of datasets and application domains. To this end, we investigate and ultimately improve ImageNet as a basis for deriving such architectures. We conduct an extensive empirical study for which we train $500$ CNN architectures, sampled from the broad AnyNetX design space, on ImageNet as well as $8$ additional well known image classification benchmark datasets from a diverse array of application domains. We observe that the performances of the architectures are highly dataset dependent. Some datasets even exhibit a negative error correlation with ImageNet across all architectures. We show how to significantly increase these correlations by utilizing ImageNet subsets restricted to fewer classes. These contributions can have a profound impact on the way we design future CNN architectures and help alleviate the tilt we see currently in our community with respect to over-reliance on one dataset.

cs.CV↗

arXiv2023

Images as Weight Matrices: Sequential Image Generation Through Synaptic Learning Rules

Work on fast weight programmers has demonstrated the effectiveness of key/value outer product-based learning rules for sequentially generating a weight matrix (WM) of a neural net (NN) by another NN or itself. However, the weight generation steps are typically not visually interpretable by humans, because the contents stored in the WM of an NN are not. Here we apply the same principle to generate natural images. The resulting fast weight painters (FPAs) learn to execute sequences of delta learning rules to sequentially generate images as sums of outer products of self-invented keys and values, one rank at a time, as if each image was a WM of an NN. We train our FPAs in the generative adversarial networks framework, and evaluate on various image datasets. We show how these generic learning rules can generate images with respectable visual quality without any explicit inductive bias for images. While the performance largely lags behind the one of specialised state-of-the-art image generators, our approach allows for visualising how synaptic learning rules iteratively produce complex connection patterns, yielding human-interpretable meaningful images. Finally, we also show that an additional convolutional U-Net (now popular in diffusion models) at the output of an FPA can learn one-step "denoising" of FPA-generated images to enhance their quality. Our code is public.

cs.CV↗

arXiv2023

On Narrative Information and the Distillation of Stories

The act of telling stories is a fundamental part of what it means to be human. This work introduces the concept of narrative information, which we define to be the overlap in information space between a story and the items that compose the story. Using contrastive learning methods, we show how modern artificial neural networks can be leveraged to distill stories and extract a representation of the narrative information. We then demonstrate how evolutionary algorithms can leverage this to extract a set of narrative templates and how these templates -- in tandem with a novel curve-fitting algorithm we introduce -- can reorder music albums to automatically induce stories in them. In the process of doing so, we give strong statistical evidence that these narrative information templates are present in existing albums. While we experiment only with music albums here, the premises of our work extend to any form of (largely) independent media.

cs.CL↗

arXiv2022

Learning One Abstract Bit at a Time Through Self-Invented Experiments Encoded as Neural Networks

There are two important things in science: (A) Finding answers to given questions, and (B) Coming up with good questions. Our artificial scientists not only learn to answer given questions, but also continually invent new questions, by proposing hypotheses to be verified or falsified through potentially complex and time-consuming experiments, including thought experiments akin to those of mathematicians. While an artificial scientist expands its knowledge, it remains biased towards the simplest, least costly experiments that still have surprising outcomes, until they become boring. We present an empirical analysis of the automatic generation of interesting experiments. In the first setting, we investigate self-invented experiments in a reinforcement-providing environment and show that they lead to effective exploration. In the second setting, pure thought experiments are implemented as the weights of recurrent neural networks generated by a neural experiment generator. Initially interesting thought experiments may become boring over time.

cs.LG↗

arXiv2022

Eliminating Meta Optimization Through Self-Referential Meta Learning

Meta Learning automates the search for learning algorithms. At the same time, it creates a dependency on human engineering on the meta-level, where meta learning algorithms need to be designed. In this paper, we investigate self-referential meta learning systems that modify themselves without the need for explicit meta optimization. We discuss the relationship of such systems to in-context and memory-based meta learning and show that self-referential neural networks require functionality to be reused in the form of parameter sharing. Finally, we propose fitness monotonic execution (FME), a simple approach to avoid explicit meta optimization. A neural network self-modifies to solve bandit and classic control tasks, improves its self-modifications, and learns how to learn, purely by assigning more computational resources to better performing solutions.

cs.LG↗

arXiv2022

Unsupervised Learning of Temporal Abstractions with Slot-based Transformers

The discovery of reusable sub-routines simplifies decision-making and planning in complex reinforcement learning problems. Previous approaches propose to learn such temporal abstractions in a purely unsupervised fashion through observing state-action trajectories gathered from executing a policy. However, a current limitation is that they process each trajectory in an entirely sequential manner, which prevents them from revising earlier decisions about sub-routine boundary points in light of new incoming information. In this work we propose SloTTAr, a fully parallel approach that integrates sequence processing Transformers with a Slot Attention module and adaptive computation for learning about the number of such sub-routines in an unsupervised fashion. We demonstrate how SloTTAr is capable of outperforming strong baselines in terms of boundary point discovery, even for sequences containing variable amounts of sub-routines, while being up to 7x faster to train on existing benchmarks.

cs.LG↗

arXiv2022

Exploring through Random Curiosity with General Value Functions

Efficient exploration in reinforcement learning is a challenging problem commonly addressed through intrinsic rewards. Recent prominent approaches are based on state novelty or variants of artificial curiosity. However, directly applying them to partially observable environments can be ineffective and lead to premature dissipation of intrinsic rewards. Here we propose random curiosity with general value functions (RC-GVF), a novel intrinsic reward function that draws upon connections between these distinct approaches. Instead of using only the current observation's novelty or a curiosity bonus for failing to predict precise environment dynamics, RC-GVF derives intrinsic rewards through predicting temporally extended general value functions. We demonstrate that this improves exploration in a hard-exploration diabolical lock problem. Furthermore, RC-GVF significantly outperforms previous methods in the absence of ground-truth episodic counts in the partially observable MiniGrid environments. Panoramic observations on MiniGrid further boost RC-GVF's performance such that it is competitive to baselines exploiting privileged information in form of episodic counts.

cs.LG↗

arXiv2022

Learning to Control Rapidly Changing Synaptic Connections: An Alternative Type of Memory in Sequence Processing Artificial Neural Networks

Short-term memory in standard, general-purpose, sequence-processing recurrent neural networks (RNNs) is stored as activations of nodes or "neurons." Generalising feedforward NNs to such RNNs is mathematically straightforward and natural, and even historical: already in 1943, McCulloch and Pitts proposed this as a surrogate to "synaptic modifications" (in effect, generalising the Lenz-Ising model, the first non-sequence processing RNN architecture of the 1920s). A lesser known alternative approach to storing short-term memory in "synaptic connections" -- by parameterising and controlling the dynamics of a context-sensitive time-varying weight matrix through another NN -- yields another "natural" type of short-term memory in sequence processing NNs: the Fast Weight Programmers (FWPs) of the early 1990s. FWPs have seen a recent revival as generic sequence processors, achieving competitive performance across various tasks. They are formally closely related to the now popular Transformers. Here we present them in the context of artificial NNs as an abstraction of biological NNs -- a perspective that has not been stressed enough in previous FWP work. We first review aspects of FWPs for pedagogical purposes, then discuss connections to related works motivated by insights from neuroscience.

cs.NE↗

arXiv2022

Neural Differential Equations for Learning to Program Neural Nets Through Continuous Learning Rules

Neural ordinary differential equations (ODEs) have attracted much attention as continuous-time counterparts of deep residual neural networks (NNs), and numerous extensions for recurrent NNs have been proposed. Since the 1980s, ODEs have also been used to derive theoretical results for NN learning rules, e.g., the famous connection between Oja's rule and principal component analysis. Such rules are typically expressed as additive iterative update processes which have straightforward ODE counterparts. Here we introduce a novel combination of learning rules and Neural ODEs to build continuous-time sequence processing nets that learn to manipulate short-term memory in rapidly changing synaptic connections of other nets. This yields continuous-time counterparts of Fast Weight Programmers and linear Transformers. Our novel models outperform the best existing Neural Controlled Differential Equation based models on various time series classification tasks, while also addressing their fundamental scalability limitations. Our code is public.

cs.LG↗

arXiv2022

CTL++: Evaluating Generalization on Never-Seen Compositional Patterns of Known Functions, and Compatibility of Neural Representations

Well-designed diagnostic tasks have played a key role in studying the failure of neural nets (NNs) to generalize systematically. Famous examples include SCAN and Compositional Table Lookup (CTL). Here we introduce CTL++, a new diagnostic dataset based on compositions of unary symbolic functions. While the original CTL is used to test length generalization or productivity, CTL++ is designed to test systematicity of NNs, that is, their capability to generalize to unseen compositions of known functions. CTL++ splits functions into groups and tests performance on group elements composed in a way not seen during training. We show that recent CTL-solving Transformer variants fail on CTL++. The simplicity of the task design allows for fine-grained control of task difficulty, as well as many insightful analyses. For example, we measure how much overlap between groups is needed by tested NNs for learning to compose. We also visualize how learned symbol representations in outputs of functions from different groups are compatible in case of success but not in case of failure. These results provide insights into failure cases reported on more complex compositions in the natural language domain. Our code is public.

cs.LG↗

arXiv2022

Learning to Generalize with Object-centric Agents in the Open World Survival Game Crafter

Reinforcement learning agents must generalize beyond their training experience. Prior work has focused mostly on identical training and evaluation environments. Starting from the recently introduced Crafter benchmark, a 2D open world survival game, we introduce a new set of environments suitable for evaluating some agent's ability to generalize on previously unseen (numbers of) objects and to adapt quickly (meta-learning). In Crafter, the agents are evaluated by the number of unlocked achievements (such as collecting resources) when trained for 1M steps. We show that current agents struggle to generalize, and introduce novel object-centric agents that improve over strong baselines. We also provide critical insights of general interest for future work on Crafter through several experiments. We show that careful hyper-parameter tuning improves the PPO baseline agent by a large margin and that even feedforward agents can unlock almost all achievements by relying on the inventory display. We achieve new state-of-the-art performance on the original Crafter environment. Additionally, when trained beyond 1M steps, our tuned agents can unlock almost all achievements. We show that the recurrent PPO agents improve over feedforward ones, even with the inventory information removed. We introduce CrafterOOD, a set of 15 new environments that evaluate OOD generalization. On CrafterOOD, we show that the current agents fail to generalize, whereas our novel object-centric agents achieve state-of-the-art OOD generalization while also being interpretable. Our code is public.

cs.LG↗

arXiv2022

General Policy Evaluation and Improvement by Learning to Identify Few But Crucial States

Learning to evaluate and improve policies is a core problem of Reinforcement Learning (RL). Traditional RL algorithms learn a value function defined for a single policy. A recently explored competitive alternative is to learn a single value function for many policies. Here we combine the actor-critic architecture of Parameter-Based Value Functions and the policy embedding of Policy Evaluation Networks to learn a single value function for evaluating (and thus helping to improve) any policy represented by a deep neural network (NN). The method yields competitive experimental results. In continuous control problems with infinitely many states, our value function minimizes its prediction error by simultaneously learning a small set of `probing states' and a mapping from actions produced in probing states to the policy's return. The method extracts crucial abstract knowledge about the environment in form of very few states sufficient to fully specify the behavior of many policies. A policy improves solely by changing actions in probing states, following the gradient of the value function's predictions. Surprisingly, it is possible to clone the behavior of a near-optimal policy in Swimmer-v3 and Hopper-v3 environments only by knowing how to act in 3 and 5 such learned states, respectively. Remarkably, our value function trained to evaluate NN policies is also invariant to changes of the policy architecture: we show that it allows for zero-shot learning of linear policies competitive with the best policy seen during training. Our code is public.

cs.LG↗

arXiv2022

Goal-Conditioned Generators of Deep Policies

Goal-conditioned Reinforcement Learning (RL) aims at learning optimal policies, given goals encoded in special command inputs. Here we study goal-conditioned neural nets (NNs) that learn to generate deep NN policies in form of context-specific weight matrices, similar to Fast Weight Programmers and other methods from the 1990s. Using context commands of the form "generate a policy that achieves a desired expected return," our NN generators combine powerful exploration of parameter space with generalization across commands to iteratively find better and better policies. A form of weight-sharing HyperNetworks and policy embeddings scales our method to generate deep NNs. Experiments show how a single learned policy generator can produce policies that achieve any return seen during training. Finally, we evaluate our algorithm on a set of continuous control tasks where it exhibits competitive performance. Our code is public.

cs.LG↗

arXiv2022

A Modern Self-Referential Weight Matrix That Learns to Modify Itself

The weight matrix (WM) of a neural network (NN) is its program. The programs of many traditional NNs are learned through gradient descent in some error function, then remain fixed. The WM of a self-referential NN, however, can keep rapidly modifying all of itself during runtime. In principle, such NNs can meta-learn to learn, and meta-meta-learn to meta-learn to learn, and so on, in the sense of recursive self-improvement. While NN architectures potentially capable of implementing such behaviour have been proposed since the '90s, there have been few if any practical studies. Here we revisit such NNs, building upon recent successes of fast weight programmers and closely related linear Transformers. We propose a scalable self-referential WM (SRWM) that learns to use outer products and the delta update rule to modify itself. We evaluate our SRWM in supervised few-shot learning and in multi-task reinforcement learning with procedurally generated game environments. Our experiments demonstrate both practical applicability and competitive performance of the proposed SRWM. Our code is public.

cs.LG↗

arXiv2022

The Dual Form of Neural Networks Revisited: Connecting Test Time Predictions to Training Patterns via Spotlights of Attention

Linear layers in neural networks (NNs) trained by gradient descent can be expressed as a key-value memory system which stores all training datapoints and the initial weights, and produces outputs using unnormalised dot attention over the entire training experience. While this has been technically known since the 1960s, no prior work has effectively studied the operations of NNs in such a form, presumably due to prohibitive time and space complexities and impractical model sizes, all of them growing linearly with the number of training patterns which may get very large. However, this dual formulation offers a possibility of directly visualising how an NN makes use of training patterns at test time, by examining the corresponding attention weights. We conduct experiments on small scale supervised image classification tasks in single-task, multi-task, and continual learning settings, as well as language modelling, and discuss potentials and limits of this view for better understanding and interpreting how NNs exploit training patterns. Our code is public.

cs.LG↗

arXiv2022

Upside-Down Reinforcement Learning Can Diverge in Stochastic Environments With Episodic Resets

Upside-Down Reinforcement Learning (UDRL) is an approach for solving RL problems that does not require value functions and uses only supervised learning, where the targets for given inputs in a dataset do not change over time. Ghosh et al. proved that Goal-Conditional Supervised Learning (GCSL) -- which can be viewed as a simplified version of UDRL -- optimizes a lower bound on goal-reaching performance. This raises expectations that such algorithms may enjoy guaranteed convergence to the optimal policy in arbitrary environments, similar to certain well-known traditional RL algorithms. Here we show that for a specific episodic UDRL algorithm (eUDRL, including GCSL), this is not the case, and give the causes of this limitation. To do so, we first introduce a helpful rewrite of eUDRL as a recursive policy update. This formulation helps to disprove its convergence to the optimal policy for a wide class of stochastic environments. Finally, we provide a concrete example of a very simple environment where eUDRL diverges. Since the primary aim of this paper is to present a negative result, and the best counterexamples are the simplest ones, we restrict all discussions to finite (discrete) environments, ignoring issues of function approximation and limited sample size.

stat.ML↗