With the advent of 5G, the internet has entered a new video-centric era. From short-video platforms like TikTok to long-video platforms like Bilibili, online video services are reshaping user consumption habits. Adaptive Bitrate (ABR) control is widely recognized as a critical factor influencing Quality of Experience (QoE). Recent learning-based ABR methods have attracted increasing attention. However, most of them rely on limited network trace sets during training and overlook the wide-distribution characteristics of real-world network conditions, resulting in poor generalization in out-of-distribution (OOD) scenarios. To address this limitation, we propose SABR, a training framework that combines behavior cloning (BC) pretraining with reinforcement learning (RL) fine-tuning. We also introduce benchmarks, ABRBench-3G and ABRBench-4G+, which provide wide-coverage training traces and dedicated OOD test sets for assessing robustness to unseen network conditions. Experimental results demonstrate that SABR achieves the best average rank compared with Pensieve, Comyco, and NetLLM across the proposed benchmarks. These results indicate that SABR enables more stable learning across wide distributions and improves generalization to unseen network conditions.

Fingerprint Presentation Attack Detection (PAD) is crucial for ensuring the reliability of fingerprint recognition systems. Mainstream methods typically frame PAD as a binary classification task. Although they achieve commendable results, these approaches often fail on unseen spoofs. This is because their decision boundaries rely heavily on trainset spoofs, leading to poor generalization. To address this issue, we reformulate PAD as an anomaly detection problem and propose the Fixed-Prototype PAD method. First, we introduce Fixed-Prototype Contrastive Learning (FPCL) for extracting PAD features. FPCL clusters the features of live fingerprints near a fixed prototype while pushing the features of spoof fingerprints away. We then design a Fixed-Prototype Anomaly Detection method to replace binary classification for decision-making. Specifically, this involves comparing the distance between test fingerprint features and the fixed live prototype to detect spoofs. Our method eliminates the dependency on the trainset spoofs during decision-making, thereby enhancing generalization capability. Experimental results demonstrate that our approach achieves state-of-the-art performance on the LivDet 2019 and LivDet 2021 benchmarks.

Automatic counterspeech generation methods have been developed to assist efforts in combating hate speech.Existing research focuses on generating counterspeech with linguistic attributes such as being polite, informative, and intent-driven.However, the real impact of counterspeech in online environments is seldom considered.This study aims to develop methods for generating counterspeech constrained by conversation outcomes and evaluate their effectiveness.We experiment with large language models (LLMs) to incorporate into the text generation process two desired conversation outcomes: low conversation incivility and nonhateful hater reentry.Specifically, we experiment with instruction prompts, LLM finetuning, and LLM reinforcement learning (RL).Evaluation results show that our methods effectively steer the generation of counterspeech towards the desired outcomes.Our analyses, however, show that there are differences in the quality and style depending on the model.

We present a model that is capable of synchronizing expressive gestures with speech. The model, implemented on a Honda humanoid robot, can generate a full range of gesture types, such as emblems, iconic and metaphoric gestures, deictic pointing and beat gestures. Arbitrary input text is analyzed with a part-of-speech tagger and a text-to-speech engine for timing information of spoken words. In addition, style tags can be optionally added to specify the level of excitement or topic changes. The text, combined with any tags, is then processed by several grammars, one for each gesture type to produce several candidate gestures for each word of the text. The model then selects probabilistically amongst the gesture types based on the desired degree of expressivity. Once a gesture type is selected, it coincides with a particular gesture template, consisting of trajectory curves that define the gesture. Speech timing patterns and style parameters are used to modulate the shape of the curve before it sent to the whole body control system on the robot. Evaluation of the model's parameters were performed, demonstrating the ability of observers to differentiate varying levels of expressiveness, excitement and speech synchronization. Modification of gesture speed for trajectory tracking found that positive associations like happiness and excitement accompanied faster speeds, with negative associations like sadness or tiredness occurred at slower speeds.