In the next few years, applications of Generative AI are expected to revolutionize a number of different areas, ranging from science & medicine to education. The potential for these seismic changes has triggered a lively debate about potential risks and resulted in calls for tighter regulation, in particular from some of the major tech companies who are leading in AI development. While regulation is important, it is key that it does not put at risk the budding field of open-source Generative AI. We argue for the responsible open sourcing of generative AI models in the near and medium term. To set the stage, we first introduce an AI openness taxonomy system and apply it to 40 current large language models. We then outline differential benefits and risks of open versus closed source AI and present potential risk mitigation, ranging from best practices to calls for technical and scientific contributions. We hope that this report will add a much needed missing voice to the current public discourse on near to mid-term AI safety and other societal impact.

Online continual learning, the process of training models on streaming data, has gained increasing attention in recent years. However, a critical aspect often overlooked is the label delay, where new data may not be labeled due to slow and costly annotation processes. We introduce a new continual learning framework with explicit modeling of the label delay between data and label streams over time steps. In each step, the framework reveals both unlabeled data from the current time step t and labels delayed with d steps, from the time step t−d. In our extensive experiments amounting to 1060 GPU days, we show that merely augmenting the computational resources is insufficient to tackle this challenge. Our findings underline a notable performance decline when solely relying on labeled data when the label delay becomes significant. More surprisingly, when using state-of-the-art SSL and TTA techniques to utilize the newer, unlabeled data, they fail to surpass the performance of a naïve method that simply trains on the delayed supervised stream. To this end, we introduce a simple, efficient baseline that rehearses from the labeled memory samples that are most similar to the new unlabeled samples. This method bridges the accuracy gap caused by label delay without significantly increasing computational complexity. We show experimentally that our method is the least affected by the label delay factor and in some cases successfully recovers the accuracy of the non-delayed counterpart. We conduct various ablations and sensitivity experiments, demonstrating the effectiveness of our approach.

In recent years by the popularization of AI, an increasing number of enterprises deployed machine learning algorithms in real life settings. This trend shed light on leaking spots of the Deep Learning bubble, namely the catastrophic decrease in quality when the distribution of the test data shifts from the training data. It is of utmost importance that we treat the promises of novel algorithms with caution and discourage reporting near perfect experimental results by fine-tuning on fixed test sets and finding metrics that hide weak points of the proposed methods. To support the wider acceptance of computer vision solutions we share our findings through a case-study in which we built a face-recognition system from scratch using consumer grade devices only, collected a database of 100k images from 150 subjects and carried out extensive validation of the most prominent approaches in single-frame face recognition literature. We show that the reported worst-case score, 74.3% true-positive ratio drops below 46.8% on real data. To overcome this barrier, after careful error analysis of the single-frame baselines we propose a low complexity solution to cover the failure cases of the single-frame recognition methods which yields an increased stability in multi-frame recognition during test time. We validate the effectiveness of the proposal by an extensive survey among our users which evaluates to 89.5% true-positive ratio.