StreamingBench
PulseAugur coverage of StreamingBench — every cluster mentioning StreamingBench across labs, papers, and developer communities, ranked by signal.
4 day(s) with sentiment data
-
New TwinPast benchmark tests AI's historical video reasoning
Researchers have introduced TwinPast, a new benchmark designed to evaluate how well AI models utilize historical evidence in streaming video understanding. Existing benchmarks often do not adequately test this capabilit…
-
New methods improve streaming video action recognition and hallucination detection
Researchers have developed Visible-State Supervision (ViSS) and Test-Time Causal View Search (TVS) to improve action recognition in streaming video. ViSS addresses state under-specification by training on source-grounde…
-
New methods improve streaming video action recognition by separating representation and view selection
Researchers have introduced Visible-State Supervision (ViSS) and Test-Time Causal View Search (TVS) to enhance streaming video action recognition. ViSS addresses state under-specification by training on probes of event …
-
SPARE framework enhances streaming video understanding with adaptive retrieval
Researchers have developed SPARE, a novel framework designed to enhance streaming video understanding by intelligently integrating historical context. SPARE uses a lightweight gate to determine if recent observations ar…
-
DeltaStreamer agent offers human-like video understanding with change-aware memory
Researchers have introduced DeltaStreamer, an open-source agent designed for understanding streaming video content. This framework mimics human observation and investigation by employing a change-aware memory system and…
-
FreshMem introduces brain-inspired memory for streaming video LLMs
Researchers have introduced FreshMem, a novel memory network designed to enhance the streaming video understanding capabilities of Multimodal Large Language Models (MLLMs). Inspired by the human brain's memory processes…
-
Foresight architecture enables proactive perception in streaming VLMs
Researchers have developed FORESIGHT, a novel dual-stream architecture for streaming vision-language models (VLMs) that enables proactive perception without retraining. By using two Siamese LLMs with shared weights, one…
-
New frameworks enhance streaming video understanding with proactive memory and detailed evaluation
Researchers have introduced several new frameworks and benchmarks for improving streaming video understanding. OneStreamer, a 4B model, unifies perception, memory, and proactive response by creating reusable factual mem…
-
Realtime-Venus system enables full-duplex interaction with dual AI models · 2 sources tracked
Researchers have introduced Realtime-Venus, a novel full-duplex interaction system designed for more natural human-computer communication. The system utilizes two specialized 9B models: Realtime-Venus-Omni for audio-vis…
-
New AI models enhance streaming video understanding with efficient memory management · 7 sources tracked
Researchers are developing new methods for streaming video understanding, focusing on efficient memory management for multimodal large language models (MLLMs). LatentStream and ShallowStream propose novel frameworks tha…
-
StreamOPD enhances streaming video understanding with new post-training technique
Researchers have introduced StreamOPD, a novel post-training technique designed to enhance streaming video understanding. This method utilizes on-policy distillation with verifiable rewards and a spatio-temporal cue-gat…
-
StreamTTT enhances streaming VLMs with real-time perception and long-term memory
Researchers have developed StreamTTT, a novel approach to enhance streaming Vision-Language Models (VLMs) by balancing real-time perception with long-term memory. Unlike previous methods that sacrifice recall for immedi…
-
VideoLLMs can now watch and think simultaneously with new VST paradigm
Researchers have introduced Video Streaming Thinking (VST), a new paradigm designed to enable online Video Large Language Models (VideoLLMs) to process and reason about video content simultaneously. This approach aims t…
-
FOLIO system enhances streaming video understanding with focused semantic memory
Researchers have introduced FOLIO, a novel training-free system designed for understanding streaming video. FOLIO addresses the challenge of unbounded memory costs in continuous video streams by intelligently compressin…
-
SelectStream framework enhances streaming video AI with selective memory allocation
Researchers have developed SelectStream, a novel framework for streaming video understanding models that selectively allocates memory to address the challenge of processing continuous video data under fixed computationa…