Video data, especially long-form video, is extremely dense and high-dimensional. Text-based summaries of video content offer a way to represent query-relevant content in a much more compact manner than raw video. In addition, textual representations are easily ingested by state-of-the-art large language models (LLMs), which enable reasoning over video content to answer complex natural language queries. To solve this issue, we rely on the progressive construction of a text-based memory by a video captioner operating on shorter chunks of the video, where spatio-temporal modeling is computationally feasible. We explore ways to improve the quality of the activity log comprised solely of short video captions. Because the video captions tend to be focused on human actions, and questions may pertain to other information in the scene, we seek to enrich the memory with static scene descriptions using Vision Language Models (VLMs). Our video understanding system relies on the LaViLa video captioner in combination with a LLM to answer questions about videos. We first explored different ways of partitioning the video into meaningful segments such that the textual descriptions more accurately reflect the structure of the video content. Furthermore, we incorporated static scene descriptions into the captioning pipeline using LLaVA VLM, resulting in a more detailed and complete caption log and expanding the space of questions that are answerable from the textual memory. Finally, we have successfully fine-tuned the LaViLa video captioner to produce both action and scene captions, significantly improving the efficiency of the captioning pipeline compared to using separate captioning models for the two tasks. Our model, controllable hybrid captioner, can alternate between different types of captions according to special input tokens that signals scene changes detected in the video.

The computational sparsity of Mixture-of-Experts (MoE) models enables sub-linear growth in compute cost as model size increases, thus offering a scalable path to training massive neural networks. However, existing implementations suffer from low GPU utilization, significant latency overhead, and a fundamental inability to leverage task locality, primarily due to CPU-managed scheduling, host-initiated communication, and frequent kernel launches. To overcome these limitations, we develop FlashMoE, a fully GPU-resident MoE operator that fuses expert computation and inter-GPU communication into a single persistent GPU kernel. FlashMoE enables fine-grained pipelining of dispatch, compute, and combine phases, eliminating launch overheads and reducing idle gaps. Unlike existing work, FlashMoE eliminates bulk-synchronous collectives for one-sided, device-initiated, inter-GPU (R)DMA transfers, thereby unlocking payload efficiency by eliminating bloated or redundant network payloads in sparsely activated layers. When evaluated on an 8-H100 GPU node with MoE models comprising up to 128 experts and 16K token sequences, FlashMoE achieves up to 9x higher GPU utilization, 6x lower latency, 5.7x higher throughput, and 4x better overlap efficiency compared to state-of-the-art baselines, despite using FP32, whereas the baselines use FP16. FlashMoE shows that principled GPU kernel-hardware co-design is key to unlocking the performance ceiling of large-scale distributed ML. We provide code at https://github.com/osayamenja/FlashMoE.

The collaborative paradigm of large and small language models (LMs) effectively balances performance and cost, yet its pivotal challenge lies in precisely pinpointing the moment of invocation when hallucinations arise in small LMs. Previous optimization efforts primarily focused on post-processing techniques, which were separate from the reasoning process of LMs, resulting in high computational costs and limited effectiveness. In this paper, we propose a practical invocation evaluation metric called AttenHScore, which calculates the accumulation and propagation of hallucinations during the generation process of small LMs, continuously amplifying potential reasoning errors. By dynamically adjusting the detection threshold, we achieve more accurate real-time invocation of large LMs. Additionally, considering the limited reasoning capacity of small LMs, we leverage uncertainty-aware knowledge reorganization to assist them better capture critical information from different text chunks. Extensive experiments reveal that our AttenHScore outperforms most baselines in enhancing real-time hallucination detection capabilities across multiple QA datasets, especially when addressing complex queries. Moreover, our strategies eliminate the need for additional model training and display flexibility in adapting to various transformer-based LMs.

Large Language Models (LLMs) exhibit pronounced memory-bound characteristics during inference due to High Bandwidth Memory (HBM) bandwidth constraints. In this paper, we propose an L2 Cache-oriented asynchronous KV Cache prefetching method to break through the memory bandwidth bottleneck in LLM inference through computation-load overlap. By strategically scheduling idle memory bandwidth during active computation windows, our method proactively prefetches required KV Cache into GPU L2 cache, enabling high-speed L2 cache hits for subsequent accesses and effectively hiding HBM access latency within computational cycles. Extensive experiments on NVIDIA H20 GPUs demonstrate that the proposed method achieves 2.15x improvement in attention kernel efficiency and up to 1.97x end-to-end throughput enhancement, surpassing state-of-the-art baseline FlashAttention-3. Notably, our solution maintains orthogonality to existing optimization techniques and can be integrated with current inference frameworks, providing a scalable latency-hiding solution for next-generation LLM inference engines.

The ability to reason lies at the core of artificial intelligence (AI), and challenging problems usually call for deeper and longer reasoning to tackle. A crucial question about AI reasoning is whether models can extrapolate learned reasoning patterns to solve harder tasks with longer chain-of-thought (CoT). In this work, we present a theoretical analysis of transformers learning on synthetic state-tracking tasks with gradient descent. We mathematically prove how the algebraic structure of state-tracking problems governs the degree of extrapolation of the learned CoT. Specifically, our theory characterizes the length generalization of transformers through the mechanism of attention concentration, linking the retrieval robustness of the attention layer to the state-tracking task structure of long-context reasoning. Moreover, for transformers with limited reasoning length, we prove that a recursive self-training scheme can progressively extend the range of solvable problem lengths. To our knowledge, we provide the first optimization guarantee that constant-depth transformers provably learn $\mathsf{NC}^1$-complete problems with CoT, significantly going beyond prior art confined in $\mathsf{TC}^0$, unless the widely held conjecture $\mathsf{TC}^0 \neq \mathsf{NC}^1$ fails. Finally, we present a broad set of experiments supporting our theoretical results, confirming the length generalization behaviors and the mechanism of attention concentration.

The rapid adoption of large language models (LLMs), such as GPT-4 and Claude 3.5, underscores the need to distinguish LLM-generated text from human-written content to mitigate the spread of misinformation and misuse in education. One promising approach to address this issue is the watermark technique, which embeds subtle statistical signals into LLM-generated text to enable reliable identification. In this paper, we first generalize the likelihood-based LLM detection method of a previous study by introducing a flexible weighted formulation, and further adapt this approach to the inverse transform sampling method. Moving beyond watermark detection, we extend this adaptive detection strategy to tackle the more challenging problem of segmenting a given text into watermarked and non-watermarked substrings. In contrast to the approach in a previous study, which relies on accurate estimation of next-token probabilities that are highly sensitive to prompt estimation, our proposed framework removes the need for precise prompt estimation. Extensive numerical experiments demonstrate that the proposed methodology is both effective and robust in accurately segmenting texts containing a mixture of watermarked and non-watermarked content.

Agentic language models compose multi step reasoning chains, yet intermediate steps can be corrupted by inconsistent context, retrieval errors, or adversarial inputs, which makes post hoc evaluation too late because errors propagate before detection. We introduce a diagnostic that requires no additional training and uses only the forward pass to emit a binary accept or reject signal during agent execution. The method analyzes token graphs induced by attention and computes two spectral statistics in early layers, namely the high frequency energy ratio and spectral entropy. We formalize these signals, establish invariances, and provide finite sample estimators with uncertainty quantification. Under a two regime mixture assumption with a monotone likelihood ratio property, we show that a single threshold on the high frequency energy ratio is optimal in the Bayes sense for detecting context inconsistency. Empirically, the high frequency energy ratio exhibits robust bimodality during context verification across multiple model families, which enables gating decisions with overhead below one millisecond on our hardware and configurations. We demonstrate integration into retrieval augmented agent pipelines and discuss deployment as an inline safety monitor. The approach detects contamination while the model is still processing the text, before errors commit to the reasoning chain.

The company is pursuing aggressive legal tactics against its opponents. On September 12, Jay Edelson received what he expected to be a standard legal document. Edelson is a lawyer representing the parents of Adam Raine; they are suing OpenAI, alleging that their 16-year-old son took his life at the encouragement of ChatGPT. OpenAI's lawyers had some inquiries for the opposing counsel, which is normal. For instance, they requested information about therapy Raine may have received, and Edelson complied.

When you purchase through links in our articles, we may earn a small commission. Google Drive's shared files and folders can now have permissions auto-expire Google Drive users can now grant temporary access to a file or folder, which is automatically revoked after a specified date. Google is now making it possible to set expiration dates on shared files and folders in Google Drive, reports Digital Trends . The feature applies to both shared drives and "My Drive," where users can grant temporary access to a file or folder that's automatically revoked after a certain date. This also means that temporary editing rights in "My Drive" no longer disappear completely when the period expires.

AI may blunt our thinking skills - here's what you can do about it There is growing evidence that our reliance on generative AI tools is reducing our ability to think clearly and critically, but it doesn't have to be that way Socrates wasn't the greatest fan of the written word. Famous for leaving no texts to posterity, the great philosopher is said to have believed that a reliance on writing destroys the memory and weakens the mind . Some 2400 years later, Socrates's fears seem misplaced - particularly in light of evidence that writing things down improves memory formation . A growing number of psychologists, neuroscientists and philosophers worry that ChatGPT and similar generative AI tools will chip away at our powers of information recall and blunt our capacity for clear reasoning. What's more, while Socrates relied on clever rhetoric to make his argument, these researchers are grounding theirs in empirical data.