3911

Agents relying purely on imitation learning (IL) or reinforcement learning (RL) often struggle to avoid safety-critical errors during exploration. Existing RL approaches for environments such as chess require hundreds of thousands of episodes and substantial computational resources to converge. Unlike RL, IL enables efficient policy learning from expert demonstrations without requiring costly trial-and-error interaction, making it well-suited for complex domains like chess. In addition, IL can capture nuanced, human-aligned behaviors such as strategic patterns and stylistic preferences. However, IL-based agents may inherit biases from training data and lack safeguards against rare but critical mistakes, such as tactical blunders that can decisively impact game outcomes. In this work, we propose Oracle-Guided Soft Shielding (OGSS), a simple yet effective framework for safer move prediction, enabling safe exploration by learning a probabilistic safety model from oracle feedback in an imitation learning setting. In the domain of chess, we employ a two-model architecture: a move prediction model predicts strong moves from expert games, and a blunder model, trained on Stockfish evaluations, estimates tactical risk. At inference, the agent scores candidate moves using a utility function that combines move likelihood and blunder probability, enabling safe yet competitive game-play. When compared to other prominent methods such as action pruning, SafeDAgger, and uncertainty-based sampling, our results demonstrate that OGSS variants maintain a lower blunder rate even as the agent’s exploration ratio is increased by several folds, highlighting its ability to support broader exploration without compromising tactical soundness.

Traditional per-title encoding schemes aim to optimize encoding resolutions to deliver the highest perceptual quality for each representation. XPSNR is observed to correlate better with the subjective quality of VVC-coded bitstreams. Towards this realization, we predict the average XPSNR of VVC-coded bitstreams using spatiotemporal complexity features of the video and the target encoding configuration using an XGBoost-based model. Based on the predicted XPSNR scores, we introduce a Quality-Aware Dynamic Resolution Adaptation (QADRA) framework for adaptive video streaming applications, where we determine the convex-hull online. Furthermore, keeping the encoding and decoding times within an acceptable threshold is mandatory for smooth and energy-efficient streaming. Hence, QADRA determines the encoding resolution and quantization parameter (QP) for each target bitrate by maximizing XPSNR while constraining the maximum encoding and/ or decoding time below a threshold. QADRA implements a JND-based representation elimination algorithm to remove perceptually redundant representations from the bitrate ladder. QADRA is an open-source Python-based framework published under the GNU GPLv3 license.

Conventional per-title encoding schemes strive to optimize encoding resolutions to deliver the utmost perceptual quality for each bitrate ladder representation. Nevertheless, maintaining encoding time within an acceptable threshold is equally imperative in online streaming applications. Further-more, modern client devices are equipped with the capability for fast deep-learning-based video super-resolution (VSR) techniques, enhancing the perceptual quality of the decoded bitstream. This suggests that opting for lower resolutions in representations during the encoding process can curtail the overall energy consumption without substantially compromising perceptual quality. In this context, this paper introduces a video super-resolution-based latency-aware optimized bitrate encoding scheme (ViSOR) designed for online adaptive streaming applications. ViSOR determines the encoding resolution for each target bitrate, ensuring the highest achievable perceptual quality after VSR within the bound of a maximum acceptable latency. Random forest-based prediction models are trained to predict the perceptual quality after VSR and the encoding time for each resolution using the spatiotemporal features extracted for each video segment. Experimental results show that ViSOR targeting fast super-resolution convolutional neural network (FSRCNN) achieves an overall average bitrate reduction of 24.65% and 32.70% to maintain the same PSNR and VMAF, compared to the HTTP Live Streaming (HLS) bitrate ladder encoding of 4s segments using the x265 encoder, when the maximum acceptable latency for each representation is set as two seconds. Considering a just noticeable difference (JND) of six VMAF points, the average cumulative storage consumption and encoding energy for each segment is reduced by 79.32% and 68.21%, respectively, contributing towards greener streaming.

Optimizing framerate for a given bitrate-spatial resolution pair in adaptive video streaming is essential to maintain perceptual quality while considering decoding complexity. Low framerates at low bitrates reduce compression artifacts and decrease decoding energy. We propose a novel method, Decoding-complexity aware Framerate Prediction (DECODRA), which employs a Variable Framerate Pareto-front approach to predict an optimized framerate that minimizes decoding energy under quality degradation constraints. DECODRA dynamically adjusts the framerate based on current bitrate and spatial resolution, balancing trade-offs between framerate, perceptual quality, and decoding complexity. Extensive experimentation with the Inter-4K dataset demonstrates DECODRA’s effectiveness, yielding an average decoding energy reduction of up to 13.45 %, with minimal VMAF reduction of 0.33 points at a low-quality degradation threshold, compared to the default 60 fps encoding. Even at an aggressive threshold, DECODRA achieves significant energy savings of 13.45 % while only reducing VMAF by 2.11 points. In this way, DECODRA extends mobile device battery life and reduces the energy footprint of streaming services by providing a more energy-efficient video streaming pipeline.

Adaptive live video streaming applications utilize a predefined collection of bitrate-resolution pairs, known as abitrate ladder, for simplicity and efficiency, eliminating the need for additional run-time to determine the optimal pairs during the live streaming session. These applications do not incorporate two-pass encoding methods due to increased latency. However, an optimized bitrate ladder could result in lower storage and delivery costs and improvedQuality of Experience(QoE). This paper presents a Just Noticeable Difference (JND)-aware constrained Variable Bitrate (cVBR) Two-pass Per-title encoding Scheme (JTPS) designed specifically for live video streaming. JTPS predicts a content- and JND-aware bitrate ladder using low-complexity features based onDiscrete Cosine Transform(DCT) energy and optimizes the constant rate factor (CRF) for each representation using random forest-based models. The effectiveness of JTPS is demonstrated using the open source video encoder x265, with an average bitrate reduction of 18.80% and 32.59% for the same PSNR and VMAF, respectively, compared to the standardHTTP Live Streaming(HLS) bitrate ladder using Constant Bitrate (CBR) encoding. The implementation of JTPS also resulted in a 68.96% reduction in storage space and an 18.58% reduction in encoding time for a JND of six VMAF points.

Versatile Video Coding (VVC) allows for large compression efficiency gains over its predecessor, High Efficiency Video Coding (HEVC). The added efficiency comes at the cost of increased runtime complexity, especially for encoding. It is thus highly relevant to explore all available runtime reduction options. This paper proposes a novel first pass for two-pass rate control in all-intra configuration, using low-complexity video analysis and a Random Forest (RF)-based machine learning model to derive the data required for driving the second pass. The proposed method is validated using VVenC, an open and optimized VVC encoder. Compared to the default two-pass rate control algorithm in VVenC, the proposed method achieves around 32% reduction in encoding time for the preset faster, while on average only causing 2% BD-rate increase and achieving similar rate control accuracy.

In video streaming applications, a fixed set of bitrate-resolution pairs (known as a bitrate ladder) is typically used during the entire streaming session. However, an optimized bitrate ladder per scene may result in (i) decreased storage or delivery costs or/and (ii) increased Quality of Experience. This paper introduces a Just Noticeable Difference (JND)-aware perscene bitrate ladder prediction scheme (JASLA) for adaptive video-on-demand streaming applications. JASLA predicts jointly optimized resolutions and corresponding constant rate factors (CRFs) using spatial and temporal complexity features for a given set of target bitrates for every scene, which yields an efficient constrained Variable Bitrate encoding. Moreover, bitrate-resolution pairs that yield distortion lower than one JND are eliminated. Experimental results show that, on average, JASLA yields bitrate savings of 34.42% and 42.67% to maintain the same PSNR and VMAF, respectively, compared to the reference HTTP Live Streaming (HLS) bitrate ladder Constant Bitrate encoding using x265 HEVC encoder, where the maximum resolution of streaming is Full HD (1080p). Moreover, a 54.34% average cumulative decrease in storage space is observed.

With the emergence of multiple modern video codecs, streaming service providers are forced to encode, store, and transmit bitrate ladders of multiple codecs separately, consequently suffering from additional energy costs for encoding, storage, and transmission. To tackle this issue, we introduce an online energy-efficient Multi-Codec Bitrate ladder Estimation scheme (MCBE) for adaptive video streaming applications. In MCBE, quality representations within the bitrate ladder of new-generation codecs (e.g., High Efficiency Video Coding (HEVC), Alliance for Open Media Video 1 (AV1)) that lie below the predicted rate-distortion curve of the Advanced Video Coding (AVC) codec are removed. Moreover, perceptual redundancy between representations of the bitrate ladders of the considered codecs is also minimized based on a Just Noticeable Difference (JND) threshold. Therefore, random forest-based models predict the VMAF score of bitrate ladder representations of each codec. In a live streaming session where all clients support the decoding of AVC, HEVC, and AV1, MCBE achieves impressive results, reducing cumulative encoding energy by 56.45%, storage energy usage by 94.99%, and transmission energy usage by 77.61% (considering a JND of six VMAF points). These energy reductions are in comparison to a baseline bitrate ladder encoding based on current industry practice.

The increasing demand for high-quality and low-cost video streaming services calls for the prediction of video encoding complexity. The prior prediction of video encoding complexity including encoding time and bitrate predictions are used to allocate resources and set optimized parameters for video encoding effectively. In this paper, a light-weight video encoding complexity prediction (VECP) scheme that predicts the encoding bitrate and the encoding time of video with high accuracy is proposed. Firstly, low-complexity Discrete Cosine Transform (DCT)-energy-based features, namely spatial complexity, temporal complexity, and brightness of videos are extracted, which can efficiently represent the encoding complexity of videos. The latent vectors are also extracted from a Convolutional Neural Network (CNN) with MobileNet as the backend to obtain additional features from representative frames of each video to assist the prediction process. The extreme gradient boosting (XGBoost) regression algorithm is deployed to predict video encoding complexity using the extracted features. The experimental results demonstrate that VECP predicts the encoding bitrate with an error percentage of up to 3.47% and encoding time with an error percentage of up to 2.89%, but with a significantly low overall latency of 3.5 milliseconds per frame which makes it suitable for both Video on Demand (VoD) and live streaming applications.

In live streaming applications, a fixed set of bitrate-resolution pairs (known as bitrate ladder) is generally used to avoid additional pre-processing run-time to analyze the complexity of every video content and determine the optimized bitrate ladder. Furthermore, live encoders use the fastest available preset for encoding to ensure the minimum possible latency in streaming. For live encoders, it is expected that the encoding speed is equal to the video framerate. An optimized encoding preset may result in (i) increased Quality of Experience (QoE) and (ii) improved CPU utilization while encoding. In this light, this paper introduces a Content-Adaptive encoder Preset prediction Scheme (CAPS) for adaptive live video streaming applications. In this scheme, the encoder preset is determined using Discrete Cosine Transform (DCT)-energy-based low-complexity spatial and temporal features for every video segment, the number of CPU threads allocated for each encoding instance, and the target encoding speed. Experimental results show that ChPS yields an overall quality improvement of 0.83 dB PSNR and 3.81 VMAF with the same bitrate, compared to the fastest preset encoding of the HTTP Live Streaming (HLS) bitrate ladder using $\times265$ HEVC open-source encoder. This is achieved by maintaining the desired encoding speed and reducing CPU idle time.