Efficient Self-supervised Continual Learning with Progressive Task-correlated Layer Freezing

Inspired by the success of Self-supervised learning (SSL) in learning visual representations from unlabeled data, a few recent works have studied SSL in the context of continual learning (CL), where multiple tasks are learned sequentially, giving rise to a new paradigm, namely self-supervised continual learning (SSCL). It has been shown that the SSCL outperforms supervised continual learning (SCL) as the learned representations are more informative and robust to catastrophic forgetting. However, if not designed intelligently, the training complexity of SSCL may be prohibitively high due to the inherent training cost of SSL. In this work, by investigating the task correlations in SSCL setup first, we discover an interesting phenomenon that, with the SSL-learned background model, the intermediate features are highly correlated between tasks. Based on this new finding, we propose a new SSCL method with layer-wise freezing which progressively freezes partial layers with the highest correlation ratios for each task to improve training computation efficiency and memory efficiency. Extensive experiments across multiple datasets are performed, where our proposed method shows superior performance against the SoTA SSCL methods under various SSL frameworks. For example, compared to LUMP, our method achieves 12\%/14\%/12\% GPU training time reduction, 23\%/26\%/24\% memory reduction, 35\%/34\%/33\% backward FLOPs reduction, and 1.31\%/1.98\%/1.21\% forgetting reduction without accuracy degradation on three datasets, respectively.

NASTyLinker: NIL-Aware Scalable Transformer-based Entity Linker

Entity Linking (EL) is the task of detecting mentions of entities in text and disambiguating them to a reference knowledge base. Most prevalent EL approaches assume that the reference knowledge base is complete. In practice, however, it is necessary to deal with the case of linking to an entity that is not contained in the knowledge base (NIL entity). Recent works have shown that, instead of focusing only on affinities between mentions and entities, considering inter-mention affinities can be used to represent NIL entities by producing clusters of mentions. At the same time, inter-mention affinities can help to substantially improve linking performance for known entities. With NASTyLinker, we introduce an EL approach that is aware of NIL entities and produces corresponding mention clusters while maintaining high linking performance for known entities. The approach clusters mentions and entities based on dense representations from Transformers and resolves conflicts (if more than one entity is assigned to a cluster) by computing transitive mention-entity affinities. We show the effectiveness and scalability of NASTyLinker on NILK, a dataset that is explicitly constructed to evaluate EL with respect to NIL entities. Further, we apply the presented approach to an actual EL task, namely to knowledge graph population by linking entities in Wikipedia listings, and provide an analysis of the outcome.

Exponential Hardness of Reinforcement Learning with Linear Function Approximation

A fundamental question in reinforcement learning theory is: suppose the optimal value functions are linear in given features, can we learn them efficiently? This problem's counterpart in supervised learning, linear regression, can be solved both statistically and computationally efficiently. Therefore, it was quite surprising when a recent work \cite{kane2022computational} showed a computational-statistical gap for linear reinforcement learning: even though there are polynomial sample-complexity algorithms, unless NP = RP, there are no polynomial time algorithms for this setting. In this work, we build on their result to show a computational lower bound, which is exponential in feature dimension and horizon, for linear reinforcement learning under the Randomized Exponential Time Hypothesis. To prove this we build a round-based game where in each round the learner is searching for an unknown vector in a unit hypercube. The rewards in this game are chosen such that if the learner achieves large reward, then the learner's actions can be used to simulate solving a variant of 3-SAT, where (a) each variable shows up in a bounded number of clauses (b) if an instance has no solutions then it also has no solutions that satisfy more than (1-$\epsilon$)-fraction of clauses. We use standard reductions to show this 3-SAT variant is approximately as hard as 3-SAT. Finally, we also show a lower bound optimized for horizon dependence that almost matches the best known upper bound of $\exp(\sqrt{H})$.

Inaccurate Label Distribution Learning

Label distribution learning (LDL) trains a model to predict the relevance of a set of labels (called label distribution (LD)) to an instance. The previous LDL methods all assumed the LDs of the training instances are accurate. However, annotating highly accurate LDs for training instances is time-consuming and very expensive, and in reality the collected LD is usually inaccurate and disturbed by annotating errors. For the first time, this paper investigates the problem of inaccurate LDL, i.e., developing an LDL model with noisy LDs. Specifically, we assume the noisy LD matrix is the linear combination of an ideal LD matrix and a sparse noisy matrix. Accordingly, inaccurate LDL becomes an inverse problem, i.e., recovering the ideal LD and noise matrix from the inaccurate LDs. To this end, we assume the ideal LD matrix is low-rank due to the correlation of labels. Besides, we use the local geometric structure of instances captured by a graph to assist the ideal LD recovery as if two instances are similar to each other, they are likely to share the same LD. The proposed model is finally formulated as a graph-regularized low-rank and sparse decomposition problem and numerically solved by the alternating direction method of multipliers. Extensive experiments demonstrate that our method can recover a relatively accurate LD from the inaccurate LD and promote the performance of different LDL methods with inaccurate LD.

CERiL: Continuous Event-based Reinforcement Learning

This paper explores the potential of event cameras to enable continuous time reinforcement learning. We formalise this problem where a continuous stream of unsynchronised observations is used to produce a corresponding stream of output actions for the environment. This lack of synchronisation enables greatly enhanced reactivity. We present a method to train on event streams derived from standard RL environments, thereby solving the proposed continuous time RL problem. The CERiL algorithm uses specialised network layers which operate directly on an event stream, rather than aggregating events into quantised image frames. We show the advantages of event streams over less-frequent RGB images. The proposed system outperforms networks typically used in RL, even succeeding at tasks which cannot be solved traditionally. We also demonstrate the value of our CERiL approach over a standard SNN baseline using event streams.

Supporting Future Electrical Utilities: Using Deep Learning Methods in EMS and DMS Algorithms

Electrical power systems are increasing in size, complexity, as well as dynamics due to the growing integration of renewable energy resources, which have sporadic power generation. This necessitates the development of near real-time power system algorithms, demanding lower computational complexity regarding the power system size. Considering the growing trend in the collection of historical measurement data and recent advances in the rapidly developing deep learning field, the main goal of this paper is to provide a review of recent deep learning-based power system monitoring and optimization algorithms. Electrical utilities can benefit from this review by re-implementing or enhancing the algorithms traditionally used in energy management systems (EMS) and distribution management systems (DMS).

Real-time Short Video Recommendation on Mobile Devices

Short video applications have attracted billions of users in recent years, fulfilling their various needs with diverse content. Users usually watch short videos on many topics on mobile devices in a short period of time, and give explicit or implicit feedback very quickly to the short videos they watch. The recommender system needs to perceive users' preferences in real-time in order to satisfy their changing interests. Traditionally, recommender systems deployed at server side return a ranked list of videos for each request from client. Thus it cannot adjust the recommendation results according to the user's real-time feedback before the next request. Due to client-server transmitting latency, it is also unable to make immediate use of users' real-time feedback. However, as users continue to watch videos and feedback, the changing context leads the ranking of the server-side recommendation system inaccurate. In this paper, we propose to deploy a short video recommendation framework on mobile devices to solve these problems. Specifically, we design and deploy a tiny on-device ranking model to enable real-time re-ranking of server-side recommendation results. We improve its prediction accuracy by exploiting users' real-time feedback of watched videos and client-specific real-time features. With more accurate predictions, we further consider interactions among candidate videos, and propose a context-aware re-ranking method based on adaptive beam search. The framework has been deployed on Kuaishou, a billion-user scale short video application, and improved effective view, like and follow by 1.28\%, 8.22\% and 13.6\% respectively.

Towards Real-Time Text2Video via CLIP-Guided, Pixel-Level Optimization

We introduce an approach to generating videos based on a series of given language descriptions. Frames of the video are generated sequentially and optimized by guidance from the CLIP image-text encoder; iterating through language descriptions, weighting the current description higher than others. As opposed to optimizing through an image generator model itself, which tends to be computationally heavy, the proposed approach computes the CLIP loss directly at the pixel level, achieving general content at a speed suitable for near real-time systems. The approach can generate videos in up to 720p resolution, variable frame-rates, and arbitrary aspect ratios at a rate of 1-2 frames per second. Please visit our website to view videos and access our open-source code: https://pschaldenbrand.github.io/text2video/ .

Impact of velocity and impact angle on football shot accuracy during fundamental trainings

The purpose of this research is to create a machine learning-based smart coaching approach for football that can replace manual analysis with real-time feedback for trainers. In-depth analysis of football player data by humans is time-consuming, error-prone, and requires a lot of effort. This exploratory study demonstrates the feasibility of using a machine learning algorithm to enhance the effectiveness of player monitoring and training. The suggested approach uses machine learning to generate analytical insights and enable long-term monitoring of player performance. In the future, machine learning could use this technique to offer constructive criticism of football players. The system incorporates a homemade ball-throwing mechanism capable of launching the ball in a variety of directions and at varying velocities. The ball kicker is equipped with a gyroscope and accelerometer sensors for measuring velocity and acceleration. The gathered data is filtered initially, and then the data that has been processed is fed into the machine-learning algorithm. The algorithm will be trained on player performance data and will be able to provide real-time feedback to coaches on player performance and potential areas for improvement. Additionally, the system will be able to track player progress over time and provide coaches with a comprehensive view of player development. The ultimate goal is to improve player performance and reduce the workload for coaches by automating the analysis process.

Robust real-time imaging through flexible multimode fibers

Conventional endoscopes comprise a bundle of optical fibers, associating one fiber for each pixel in the image. In principle, this can be reduced to a single multimode optical fiber (MMF), the width of a human hair, with one fiber spatial-mode per image pixel. However, images transmitted through a MMF emerge as unrecognisable speckle patterns due to dispersion and coupling between the spatial modes of the fiber. Furthermore, speckle patterns change as the fiber undergoes bending, making the use of MMFs in flexible imaging applications even more complicated. In this paper, we propose a real-time imaging system using flexible MMFs, but which is robust to bending. Our approach does not require access or feedback signal from the distal end of the fiber during imaging. We leverage a variational autoencoder (VAE) to reconstruct and classify images from the speckles and show that these images can still be recovered when the bend configuration of the fiber is changed to one that was not part of the training set. We utilize a MMF $300$ mm long with a 50 $\mu$m core for imaging $10\times 10$ cm objects placed approximately at $20$ cm from the fiber and the system can deal with a change in fiber bend of 50$^\circ$ and range of movement of 8 cm.