Vision-Language Models (VLMs) are increasingly required to process unbounded video streams in applications such as video-call assistants, live commentary, and embodied robots. An ideal streaming system should support proactive interaction, long-horizon memory, and real-time processing, while resting on a VLM backbone capable of handling diverse in-the-wild streaming tasks. However, existing VLMs excel at offline video understanding but fall short in streaming capabilities and lack dedicated infrastructure for streaming deployment. We address this gap on three fronts. (i) For backbone capability, we construct \textbf{Streaming-Train-248K}, a streaming dataset paired with a novel training objective for adapting VLMs to streaming interaction and understanding. (ii) For real-world deployment, we introduce \textbf{Streaming Harness}, a plug-and-play system that endows any VLM with three core abilities: proactive interaction (per-second response decisions), long-term memory (12-hour context retention), and real-time processing (sub-second latency). (iii) To drive continued community progress on streaming capabilities, we design \textbf{Streaming-Eval}, a benchmark that reflects models' c
There has been increasing interest in unifying streaming and non-streaming automatic speech recognition (ASR) models to reduce development, training, and deployment costs. We present a unified framework that trains a single end-to-end ASR model for both streaming and non-streaming applications, leveraging future context information. We propose to use dynamic right-context through the chunked attention masking in the training of zipformer-based ASR models. We demonstrate that using right-context is more effective in zipformer models compared to other conformer models due to its multi-scale nature. We analyze the effect of varying the number of right-context frames on accuracy and latency of the streaming ASR models. We use Librispeech and large in-house conversational datasets to train different versions of streaming and non-streaming models and evaluate them in a production grade server-client setup across diverse testsets of different domains. The proposed strategy reduces word error by relative 7.9\% with a small degradation in user-perceived latency. By adding more right-context frames, we are able to achieve streaming performance close to that of non-streaming models. Our appro
The unified streaming and non-streaming speech recognition model has achieved great success due to its comprehensive capabilities. In this paper, we propose to improve the accuracy of the unified model by bridging the inherent representation gap between the streaming and non-streaming modes with a contrastive objective. Specifically, the top-layer hidden representation at the same frame of the streaming and non-streaming modes are regarded as a positive pair, encouraging the representation of the streaming mode close to its non-streaming counterpart. The multiple negative samples are randomly selected from the rest frames of the same sample under the non-streaming mode. Experimental results demonstrate that the proposed method achieves consistent improvements toward the unified model in both streaming and non-streaming modes. Our method achieves CER of 4.66% in the streaming mode and CER of 4.31% in the non-streaming mode, which sets a new state-of-the-art on the AISHELL-1 benchmark.
Real-time understanding of continuous video streams is essential for interactive assistants and multimodal agents operating in dynamic environments. However, most existing video reasoning approaches follow a batch paradigm that defers reasoning until the full video context is observed, resulting in high latency and growing computational cost that are incompatible with streaming scenarios. In this paper, we introduce ThinkStream, a framework for streaming video reasoning based on a Watch--Think--Speak paradigm that enables models to incrementally update their understanding as new video observations arrive. At each step, the model performs a short reasoning update and decides whether sufficient evidence has accumulated to produce a response. To support long-horizon streaming, we propose Reasoning-Compressed Streaming Memory (RCSM), which treats intermediate reasoning traces as compact semantic memory that replaces outdated visual tokens while preserving essential context. We further train the model using a Streaming Reinforcement Learning with Verifiable Rewards scheme that aligns incremental reasoning and response timing with the requirements of streaming interaction. Experiments on
Acoustic streaming is the net time-averaged flow that results from the nonlinearities in an oscillating flow. Extensive research has sought to identify different physical mechanisms and types of acoustic streaming in systems of various geometries. While streaming in a channel maintains one of the simplest geometries, dimensional analysis of the governing equations reveals that multiple regimes of streaming may occur within a channel. In this study, a framework is developed for investigating and understanding the physical streaming regimes in a two-dimensional channel. By taking different limits of the dimensionless number ratios found within the framework, streaming models derived in previous works are recovered to demonstrate the different streaming regimes within a channel. The onset of fast streaming is then analyzed with the framework and nonlinear Reynolds numbers, which indicate whether the streaming is slow or fast, are found for the different physical streaming regimes. As a result, the framework provides a base for analyzing fast streaming in a channel and streaming in multi-scale systems while organizing previous streaming models into a physical spectrum for a channel geo
This paper presents StreamChat, a novel approach that enhances the interaction capabilities of Large Multimodal Models (LMMs) with streaming video content. In streaming interaction scenarios, existing methods rely solely on visual information available at the moment a question is posed, resulting in significant delays as the model remains unaware of subsequent changes in the streaming video. StreamChat addresses this limitation by innovatively updating the visual context at each decoding step, ensuring that the model utilizes up-to-date video content throughout the decoding process. Additionally, we introduce a flexible and efficient crossattention-based architecture to process dynamic streaming inputs while maintaining inference efficiency for streaming interactions. Furthermore, we construct a new dense instruction dataset to facilitate the training of streaming interaction models, complemented by a parallel 3D-RoPE mechanism that encodes the relative temporal information of visual and text tokens. Experimental results demonstrate that StreamChat achieves competitive performance on established image and video benchmarks and exhibits superior capabilities in streaming interaction
Sliced optimal transport (SOT), or sliced Wasserstein (SW) distance, is widely recognized for its statistical and computational scalability. In this work, we further enhance computational scalability by proposing the first method for estimating SW from sample streams, called streaming sliced Wasserstein (Stream-SW). To define Stream-SW, we first introduce a streaming estimator of the one-dimensional Wasserstein distance (1DW). Since the 1DW has a closed-form expression, given by the integral of the absolute difference between the quantile functions of the compared distributions, we leverage quantile approximation techniques for sample streams to define a streaming 1DW estimator. By applying the streaming 1DW to all projections, we obtain Stream-SW. The key advantage of Stream-SW is its low memory complexity while providing theoretical guarantees on the approximation error. We demonstrate that Stream-SW achieves a more accurate approximation of SW than random subsampling, with lower memory consumption, when comparing Gaussian distributions and mixtures of Gaussians from streaming samples. Additionally, we conduct experiments on point cloud classification, point cloud gradient flows,
Generative conversational interfaces powered by large language models (LLMs) typically stream output token-by-token at a rate determined by computational budget, often neglecting actual human reading speeds and the cognitive load associated with the content. This mismatch frequently leads to inefficient use of computational resources. For example, in cloud-based services, streaming content faster than users can read appears unnecessary, resulting in wasted computational resources and potential delays for other users, particularly during peak usage periods. To address this issue, we propose an adaptive streaming method that dynamically adjusts the pacing of LLM streaming output in real-time based on inferred cognitive load. Our approach estimates the cognitive load associated with streaming content and strategically slows down the stream during complex or information-rich segments, thereby freeing computational resources for other users. We conducted a statistical analysis and simulation based on a statistical model derived from data collected in a crowdsourced user study across various types of LLM-generated content. Our results show that this adaptive method can effectively reduce
Adaptive streaming of segmented video over HTTP typically relies on a predefined set of bitrate-resolution pairs, known as a bitrate ladder. However, fixed ladders often overlook variations in content and decoding complexities, leading to suboptimal trade-offs between encoding time, decoding efficiency, and video quality. This article introduces DQ-Ladder, a deep reinforcement learning (DRL)-based scheme for constructing time- and quality-aware bitrate ladders for adaptive video streaming applications. DQ-Ladder employs predicted decoding time, quality scores, and bitrate levels per segment as inputs to a Deep Q-Network (DQN) agent, guided by a weighted reward function of decoding time, video quality, and resolution smoothness. We leverage machine learning models to predict decoding time, bitrate level, and objective quality metrics (VMAF, XPSNR), eliminating the need for exhaustive encoding or quality metric computation. We evaluate DQ-Ladder using the Versatile Video Coding (VVC) toolchain (VVenC/VVdeC) on 750 video sequences across six Apple HLS-compliant resolutions and 41 quantization parameters. Experimental results against four baselines show that DQ-Ladder achieves BD-rate
Diffusion Large Language Models (dLLMs) offer a compelling paradigm for natural language generation, leveraging parallel decoding and bidirectional attention to achieve superior global coherence compared to autoregressive models. While recent works have accelerated inference via KV cache reuse or heuristic decoding, they overlook the intrinsic inefficiencies within the block-wise diffusion process. Specifically, they suffer from spatial redundancy by modeling informative-sparse suffix regions uniformly and temporal inefficiency by applying fixed denoising schedules across all the decoding process. To address this, we propose Streaming-dLLM, a training-free framework that streamlines inference across both spatial and temporal dimensions. Spatially, we introduce attenuation guided suffix modeling to approximate the full context by pruning redundant mask tokens. Temporally, we employ a dynamic confidence aware strategy with an early exit mechanism, allowing the model to skip unnecessary iterations for converged tokens. Extensive experiments show that Streaming-dLLM achieves up to 68.2X speedup while maintaining generation quality, highlighting its effectiveness in diffusion decoding.
Timing and burst patterns can leak through encryption, and an adaptive adversary can exploit them. This undermines metadata-only detection in a stand-alone consumer gateway. Therefore, consumer gateways need streaming intrusion detection on encrypted traffic using metadata only, under tight CPU and latency budgets. We present a streaming IDS for stand-alone gateways that instantiates a lightweight two-state unit derived from Network-Optimised Spiking (NOS) dynamics per flow, named \emph{NOS-Gate}. NOS-Gate scores fixed-length windows of metadata features and, under a $K$-of-$M$ persistence rule, triggers a reversible mitigation that temporarily reduces the flow's weight under weighted fair queueing (WFQ). We evaluate NOS-Gate under timing-controlled evasion using an executable \emph{worlds} benchmark that specifies benign device processes, auditable attacker budgets, contention structure, and packet-level WFQ replay to quantify queue impact. All methods are calibrated label-free via burn-in quantile thresholding. Across multiple reproducible worlds and malicious episodes, at an achieved $0.1\%$ false-positive operating point, NOS-Gate attains 0.952 incident recall versus 0.857 for
Streaming automatic speech recognition (ASR) models are restricted from accessing future context, which results in worse performance compared to the non-streaming models. To improve the performance of streaming ASR, knowledge distillation (KD) from the non-streaming to streaming model has been studied, mainly focusing on aligning the output token probabilities. In this paper, we propose a layer-to-layer KD from the teacher encoder to the student encoder. To ensure that features are extracted using the same context, we insert auxiliary non-streaming branches to the student and perform KD from the non-streaming teacher layer to the non-streaming auxiliary layer. We design a special KD loss that leverages the autoregressive predictive coding (APC) mechanism to encourage the streaming model to predict unseen future contexts. Experimental results show that the proposed method can significantly reduce the word error rate compared to previous token probability distillation methods.
End-to-end (E2E) automatic speech recognition (ASR) models, by now, have shown competitive performance on several benchmarks. These models are structured to either operate in streaming or non-streaming mode. This work presents cascaded encoders for building a single E2E ASR model that can operate in both these modes simultaneously. The proposed model consists of streaming and non-streaming encoders. Input features are first processed by the streaming encoder; the non-streaming encoder operates exclusively on the output of the streaming encoder. A single decoder then learns to decode either using the output of the streaming or the non-streaming encoder. Results show that this model achieves similar word error rates (WER) as a standalone streaming model when operating in streaming mode, and obtains 10% -- 27% relative improvement when operating in non-streaming mode. Our results also show that the proposed approach outperforms existing E2E two-pass models, especially on long-form speech.
Streaming recommender systems (SRSs) are widely deployed in real-world applications, where user interests shift and new items arrive over time. As a result, effectively capturing users' latest preferences is challenging, as interactions reflecting recent interests are limited and new items often lack sufficient feedback. A common solution is to enrich item representations using multimodal encoders (e.g., BERT or ViT) to extract visual and textual features. However, these encoders are pretrained on general-purpose tasks: they are not tailored to user preference modeling, and they overlook the fact that user tastes toward modality-specific features such as visual styles and textual tones can also drift over time. This presents two key challenges in streaming scenarios: the high cost of fine-tuning large multimodal encoders, and the risk of forgetting long-term user preferences due to continuous model updates. To tackle these challenges, we propose Expandable Side Mixture-of-Experts (XSMoE), a memory-efficient framework for multimodal streaming recommendation. XSMoE attaches lightweight side-tuning modules consisting of expandable expert networks to frozen pretrained encoders and incr
We introduce Delayed Streams Modeling (DSM), a flexible formulation for streaming, multimodal sequence-to-sequence learning. Sequence-to-sequence generation is often cast in an offline manner, where the model consumes the complete input sequence before generating the first output timestep. Alternatively, streaming sequence-to-sequence rely on learning a policy for choosing when to advance on the input stream, or write to the output stream. DSM instead models already time-aligned streams with a decoder-only language model. By moving the alignment to a pre-processing step,and introducing appropriate delays between streams, DSM provides streaming inference of arbitrary output sequences, from any input combination, making it applicable to many sequence-to-sequence problems. In particular, given text and audio streams, automatic speech recognition (ASR) corresponds to the text stream being delayed, while the opposite gives a text-to-speech (TTS) model. We perform extensive experiments for these two major sequence-to-sequence tasks, showing that DSM provides state-of-the-art performance and latency while supporting arbitrary long sequences, being even competitive with offline baselines.
The acoustofluidic method holds great promise for manipulating microorganisms. When exposed to the steady vortex structures of acoustic streaming flow, these microorganisms exhibit intriguing dynamic behaviors, such as hydrodynamic trapping and aggregation. To uncover the mechanisms behind these behaviors, we investigate the swimming dynamics of both passive and active particles within a two-dimensional acoustic streaming flow. By employing a theoretically calculated streaming flow field, we demonstrate the existence of stable bounded orbits for particles. Additionally, we introduce rotational diffusion and examine the distribution of particles under varying flow strengths. Our findings reveal that active particles can laterally migrate across streamlines and become trapped in stable bounded orbits closer to the vortex center, whereas passive particles are confined to movement along the streamlines. We emphasize the influence of the flow field on the distribution and trapping of active particles, identifying a flow configuration that maximizes their aggregation. These insights contribute to the manipulation of microswimmers and the development of innovative biological microfluidic
The advent of 5G has driven the demand for high-quality, low-latency live streaming. However, challenges such as managing the increased data volume, ensuring synchronization across multiple streams, and maintaining consistent quality under varying network conditions persist, particularly in real-time video streaming. To address these issues, we propose a novel framework that leverages 3D virtual environments within game engines (e.g., Unity 3D) to optimize multi-channel live streaming. Our approach consolidates multi-camera video data into a single stream using multiple virtual 3D canvases, significantly increasing channel amounts while reducing latency and enhancing user flexibility. For demonstration of our approach, we utilize the Unity 3D engine to integrate multiple video inputs into a single-channel stream, supporting one-to-many broadcasting, one-to-one video calling, and real-time control of video channels. By mapping video data onto a world-space canvas and capturing it via an in-world camera, we minimize redundant data transmission, achieving efficient, low-latency streaming. Our results demonstrate that this method outperforms some existing multi-channel live streaming s
The Click-Through Rate (CTR) prediction task is critical in industrial recommender systems, where models are usually deployed on dynamic streaming data in practical applications. Such streaming data in real-world recommender systems face many challenges, such as distribution shift, temporal non-stationarity, and systematic biases, which bring difficulties to the training and utilizing of recommendation models. However, most existing studies approach the CTR prediction as a classification task on static datasets, assuming that the train and test sets are independent and identically distributed (a.k.a, i.i.d. assumption). To bridge this gap, we formulate the CTR prediction problem in streaming scenarios as a Streaming CTR Prediction task. Accordingly, we propose dedicated benchmark settings and metrics to evaluate and analyze the performance of the models in streaming data. To better understand the differences compared to traditional CTR prediction tasks, we delve into the factors that may affect the model performance, such as parameter scale, normalization, regularization, etc. The results reveal the existence of the ''streaming learning dilemma'', whereby the same factor may have d
In streaming scenarios, models must learn continuously, adapting to concept drifts without erasing previously acquired knowledge. However, existing research communities address these challenges in isolation. Continual Learning (CL) focuses on long-term retention and mitigating catastrophic forgetting, often without strict real-time constraints. Stream Learning (SL) emphasizes rapid, efficient adaptation to high-frequency data streams, but typically neglects forgetting. Recent efforts have tried to combine these paradigms, yet no clear algorithmic overlap exists. We argue that large in-context tabular models (LTMs) provide a natural bridge for Streaming Continual Learning (SCL). In our view, unbounded streams should be summarized on-the-fly into compact sketches that can be consumed by LTMs. This recovers the classical SL motivation of compressing massive streams with fixed-size guarantees, while simultaneously aligning with the experience-replay desiderata of CL. To clarify this bridge, we show how the SL and CL communities implicitly adopt a divide-to-conquer strategy to manage the tension between plasticity (performing well on the current distribution) and stability (retaining pa
In this work we explore the latency and accuracy of keyword spotting (KWS) models in streaming and non-streaming modes on mobile phones. NN model conversion from non-streaming mode (model receives the whole input sequence and then returns the classification result) to streaming mode (model receives portion of the input sequence and classifies it incrementally) may require manual model rewriting. We address this by designing a Tensorflow/Keras based library which allows automatic conversion of non-streaming models to streaming ones with minimum effort. With this library we benchmark multiple KWS models in both streaming and non-streaming modes on mobile phones and demonstrate different tradeoffs between latency and accuracy. We also explore novel KWS models with multi-head attention which reduce the classification error over the state-of-art by 10% on Google speech commands data sets V2. The streaming library with all experiments is open-sourced.