We study the learning dynamics of self-predictive learning for reinforcement learning, a family of algorithms that learn representations by minimizing the prediction error of their own future latent representations. Despite its recent empirical success, such algorithms have an apparent defect: trivial representations (such as constants) minimize the prediction error, yet it is obviously undesirable to converge to such solutions. Our central insight is that careful designs of the optimization dynamics are critical to learning meaningful representations. We identify that a faster paced optimization of the predictor and semi-gradient updates on the representation, are crucial to preventing the representation collapse. Then in an idealized setup, we show self-predictive learning dynamics carries out spectral decomposition on the state transition matrix, effectively capturing information of the transition dynamics. Building on the theoretical insights, we propose bidirectional self-predictive learning, a novel self-predictive algorithm that learns two representations simultaneously. We examine the robustness of our theoretical insights with a number of small-scale experiments and showcase the promise of the novel representation learning algorithm with large-scale experiments.

我们考虑在以$ s $状态的地平线$ h $和$ a $ ACTIVE的偶发性，有限的，依赖于阶段的马尔可夫决策过程的环境中进行强化学习。代理商的性能是在与环境互动以$ t $插件互动后的遗憾来衡量的。我们提出了一种乐观的后验抽样算法（OPSRL），这是一种简单的后验抽样变体，仅需要许多后样品对数，$ h $，$ s $，$ a $和$ t $ a $ h $ s $ s $ a $ a $和$ t $一对。对于OPSRL，我们保证最多可容纳订单的高概率遗憾，$ \ wideTilde {\ mathcal {o}}}（\ sqrt {h^3sat}）$忽略$ \ text {poly} \ log（hsat）$项。新型的新型技术成分是线性形式的新型抗浓缩不等式，可能具有独立感兴趣。具体而言，我们将Alfers and Dinges [1984]的Beta分布的基于正常近似的下限扩展到Dirichlet分布。我们的界限匹配订单$ \ omega（\ sqrt {h^3sat}）$的下限，从而回答了Agrawal和Jia [2017b]在情节环境中提出的空旷问题。

我们提出BYOL-QUENPLORE，这是一种在视觉复杂环境中进行好奇心驱动的探索的概念上简单但一般的方法。Byol-explore通过优化潜在空间中的单个预测损失而没有其他辅助目标，从而学习了世界代表，世界动态和探索政策。我们表明，BYOL探索在DM-HARD-8中有效，DM-HARD-8是一种具有挑战性的部分可观察的连续操作硬探索基准，具有视觉富含3-D环境。在这个基准上，我们完全通过使用Byol-explore的内在奖励来纯粹通过增强外部奖励来解决大多数任务，而先前的工作只能通过人类的示威来脱颖而出。作为Byol-explore的一般性的进一步证据，我们表明它在Atari的十个最难的探索游戏中实现了超人的性能，同时设计比其他竞争力代理人要简单得多。

我们展示了一个通过ImageNet（Optip）问题，旨在研究流媒体环境中深度学习的有效性。 Imagenet是一个广泛的已知基准数据集，有助于推动和评估深度学习的最新进步。通常，深度学习方法训练在模型具有随机访问的静态数据上，使用多次通过数据集，在每个时期的训练中随机随机抽搐。这种数据访问假设在许多真实情景中不存在，其中从流中收集大规模数据并存储和访问所有数据由于存储成本和隐私问题而变得不切实际。对于拍摄，我们将ImageNet数据视为顺序到达，内存预算有限的内存预算来存储一个小的数据子集。我们观察到，在单次训练中培训一个深度网络，用于多挪训练导致预测准确性的巨大降低。我们表明，尽管对典型的连续问题设置不同，但通过支付小的记忆成本和利用为持续学习的技术来说，可以显着降低性能差距。我们建议使用参考学习资源有效的深度学习。

Pre-trained language models (PLM) have advanced the state-of-the-art across NLP applications, but lack domain-specific knowledge that does not naturally occur in pre-training data. Previous studies augmented PLMs with symbolic knowledge for different downstream NLP tasks. However, knowledge bases (KBs) utilized in these studies are usually large-scale and static, in contrast to small, domain-specific, and modifiable knowledge bases that are prominent in real-world task-oriented dialogue (TOD) systems. In this paper, we showcase the advantages of injecting domain-specific knowledge prior to fine-tuning on TOD tasks. To this end, we utilize light-weight adapters that can be easily integrated with PLMs and serve as a repository for facts learned from different KBs. To measure the efficacy of proposed knowledge injection methods, we introduce Knowledge Probing using Response Selection (KPRS) -- a probe designed specifically for TOD models. Experiments on KPRS and the response generation task show improvements of knowledge injection with adapters over strong baselines.

The current trend of applying transfer learning from CNNs trained on large datasets can be an overkill when the target application is a custom and delimited problem with enough data to train a network from scratch. On the other hand, the training of custom and lighter CNNs requires expertise, in the from-scratch case, and or high-end resources, as in the case of hardware-aware neural architecture search (HW NAS), limiting access to the technology by non-habitual NN developers. For this reason, we present Colab NAS, an affordable HW NAS technique for producing lightweight task-specific CNNs. Its novel derivative-free search strategy, inspired by Occam's razor, allows it to obtain state-of-the-art results on the Visual Wake Word dataset in just 4.5 GPU hours using free online GPU services such as Google Colaboratory and Kaggle Kernel.

In this paper we propose a general approach to define a many-valued preferential interpretation of gradual argumentation semantics. The approach allows for conditional reasoning over arguments and boolean combination of arguments, with respect to a class of gradual semantics, through the verification of graded (strict or defeasible) implications over a preferential interpretation. As a proof of concept, in the finitely-valued case, an Answer set Programming approach is proposed for conditional reasoning in a many-valued argumentation semantics of weighted argumentation graphs. The paper also develops and discusses a probabilistic semantics for gradual argumentation, which builds on the many-valued conditional semantics.

Weather forecasting centers currently rely on statistical postprocessing methods to minimize forecast error. This improves skill but can lead to predictions that violate physical principles or disregard dependencies between variables, which can be problematic for downstream applications and for the trustworthiness of postprocessing models, especially when they are based on new machine learning approaches. Building on recent advances in physics-informed machine learning, we propose to achieve physical consistency in deep learning-based postprocessing models by integrating meteorological expertise in the form of analytic equations. Applied to the post-processing of surface weather in Switzerland, we find that constraining a neural network to enforce thermodynamic state equations yields physically-consistent predictions of temperature and humidity without compromising performance. Our approach is especially advantageous when data is scarce, and our findings suggest that incorporating domain expertise into postprocessing models allows to optimize weather forecast information while satisfying application-specific requirements.

Pretrained transformer models have achieved state-of-the-art results in many tasks and benchmarks recently. Many state-of-the-art Language Models (LMs), however, do not scale well above the threshold of 512 input tokens. In specialized domains though (such as legal, scientific or biomedical), models often need to process very long text (sometimes well above 10000 tokens). Even though many efficient transformers have been proposed (such as Longformer, BigBird or FNet), so far, only very few such efficient models are available for specialized domains. Additionally, since the pretraining process is extremely costly in general - but even more so as the sequence length increases - it is often only in reach of large research labs. One way of making pretraining cheaper is the Replaced Token Detection (RTD) task, by providing more signal during training, since the loss can be computed over all tokens. In this work, we train Longformer models with the efficient RTD task on legal data to showcase that pretraining efficient LMs is possible using much less compute. We evaluate the trained models on challenging summarization tasks requiring the model to summarize long texts to show to what extent the models can achieve good performance on downstream tasks. We find that both the small and base models outperform their baselines on the in-domain BillSum and out-of-domain PubMed tasks in their respective parameter range. We publish our code and models for research purposes.

Synthetic data generation has recently gained widespread attention as a more reliable alternative to traditional data anonymization. The involved methods are originally developed for image synthesis. Hence, their application to the typically tabular and relational datasets from healthcare, finance and other industries is non-trivial. While substantial research has been devoted to the generation of realistic tabular datasets, the study of synthetic relational databases is still in its infancy. In this paper, we combine the variational autoencoder framework with graph neural networks to generate realistic synthetic relational databases. We then apply the obtained method to two publicly available databases in computational experiments. The results indicate that real databases' structures are accurately preserved in the resulting synthetic datasets, even for large datasets with advanced data types.