Analyzing unstructured data has been a persistent challenge in data processing. Large Language Models (LLMs) have shown promise in this regard, leading to recent proposals for declarative frameworks for LLM-powered processing of unstructured data. However, these frameworks focus on reducing cost when executing user-specified operations using LLMs, rather than improving accuracy, executing most operations as-is (in a single LLM call). This is problematic for complex tasks and data, where LLM outputs for user-defined operations are often inaccurate, even with optimized prompts. For example, an LLM may struggle to identify {\em all} instances of specific clauses, like force majeure or indemnification, in lengthy legal documents, requiring decomposition of the data, the task, or both. We present DocETL, a system that optimizes complex document processing pipelines, while accounting for LLM shortcomings. DocETL offers a declarative interface for users to define such pipelines and uses an agent-based approach to automatically optimize them, leveraging novel agent-based rewrites (that we call rewrite directives), as well as an optimization and evaluation framework. We introduce (i) logical rewriting of pipelines, tailored for LLM-based tasks, (ii) an agent-guided plan evaluation mechanism that synthesizes and orchestrates task-specific validation prompts, and (iii) an optimization algorithm that efficiently finds promising plans, considering the latencies of agent-based plan generation and evaluation. Our evaluation on four different unstructured document analysis tasks demonstrates that DocETL finds plans with outputs that are 25 to 80% more accurate than well-engineered baselines, addressing a critical gap in unstructured data analysis. DocETL is open-source at docetl.org, and as of November 2024, has amassed over 1.3k GitHub Stars, with users spanning a variety of domains.

Multimodal Large Language Models (MLLMs) excel in vision--language tasks by pre-training solely on coarse-grained concept annotations (e.g., image captions). We hypothesize that integrating fine-grained concept annotations (e.g., object labels and object regions) will further improve performance, as both data granularities complement each other in terms of breadth and depth in concept representation. We introduce a new dataset featuring Multimodal Multi-Grained Concept annotations (MMGiC) for MLLMs. In constructing MMGiC, we explore the impact of different data recipes on multimodal comprehension and generation. Our analyses reveal that multi-grained concept annotations integrate and complement each other, under our structured template and a general MLLM framework. We clearly explore and demonstrate the potential of MMGiC to help MLLMs better locate and learn concepts, aligning vision and language at multiple granularities. We further validate our hypothesis by investigating the fair comparison and effective collaboration between MMGiC and image--caption data on 12 multimodal comprehension and generation benchmarks, e.g., their appropriate combination achieve 3.95% and 2.34% absolute improvements over image--caption data alone on POPE and SEED-Bench. Code, data and models will be available at https://github.com/LooperXX/MMGiC.

The Earth's weather system involves intricate weather data modalities and diverse weather understanding tasks, which hold significant value to human life. Existing data-driven models focus on single weather understanding tasks (e.g., weather forecasting). While these models have achieved promising results, they fail to tackle various complex tasks within a single and unified model. Moreover, the paradigm that relies on limited real observations for a single scenario hinders the model's performance upper bound. Inspired by the in-context learning paradigm from visual foundation models and large language models, in this paper, we introduce the first generalist weather generalist foundation model (WeatherGFM) to address weather understanding tasks in a unified manner. Specifically, we first unify the representation and definition for diverse weather understanding tasks. Subsequently, we design weather prompt formats to handle different weather data modalities, including single, multiple, and temporal modalities. Finally, we adopt a visual prompting question-answering paradigm for the training of unified weather understanding tasks. Extensive experiments indicate that our WeatherGFM can effectively handle up to ten weather understanding tasks, including weather forecasting, super-resolution, weather image translation, and post-processing. Modeling Earth weather systems involves a series of complex subprocesses that are intended to transform intricate Earth observation data into applications like weather forecasting (Chen et al., 2023a; Bi et al., 2023), downscaling (Chen et al., 2022), assimilation (Huang et al., 2024), retrieval (Liu et al., 2011), and bias correction (Gong et al., 2024). During the past decade, many data-driven machine learning methods have been investigated for various weather understanding tasks and delivering desirable performance on specific tasks. For example, recent studies using large-scale training data (e.g., ERA5 reanalysis data (Hersbach et al., 2020)) have exceeded the accuracy of conventional numerical weather forecasts.

As Large Language Models (LLMs) scale to longer context windows, the computational cost of attention mechanisms, which traditionally grows quadratically with input length, presents a critical challenge for real-time and memory-constrained deployments. Existing sparse attention techniques have sought to reduce this complexity, but they often incur significant overhead or compromise accuracy, making them less practical for large contexts on mid-range hardware. In this paper, we introduce SparseAccelerate, a dynamic sparse attention method that adapts its sparsity patterns based on input characteristics, effectively flattening the attention complexity curve. Our approach is effective for input lengths starting at 16K tokens and scales efficiently up to 128K tokens on dual NVIDIA A5000 GPUs (24GB each). Experimental results show that SparseAccelerate achieves up to a 1.04x reduction in Time-To-First-Token (TTFT) latency at 32K tokens, while also providing substantial memory savings. These improvements yield practical gains for memory-intensive applications and long-context tasks that were previously infeasible with standard attention. Beyond latency reductions, SparseAccelerate fundamentally shifts the scaling trend, demonstrating the smallest TTFT growth gradient relative to context length among competing methods. Ongoing evaluations on diverse benchmarks confirm its scalability, positioning SparseAccelerate as a critical advancement toward efficient, real-time, and large-context LLM inference on accessible hardware.

Abstract--Human-interactive robotic systems, particularly autonomous vehicles (AVs), must effectively integrate human instructions into their motion planning. This paper introduces doScenes, a novel dataset designed to facilitate research on human-vehicle instruction interactions, focusing on short-term directives that directly influence vehicle motion. Unlike existing datasets that focus on ranking or scenelevel reasoning, doScenes emphasizes actionable directives tied to static and dynamic scene objects. This framework addresses limitations in prior research, such as reliance on simulated data or predefined action sets, by supporting nuanced and flexible responses in real-world scenarios. This work lays the foundation for developing learning strategies that seamlessly integrate human instructions into autonomous systems, advancing safe and effective human-vehicle collaboration. In the doScenes dataset, we augment each clip of temporal data with an instruction and a tag to indicate the instruction's referentiality.

Software engineers frequently grapple with the challenge of accessing disparate documentation and telemetry data, including Troubleshooting Guides (TSGs), incident reports, code repositories, and various internal tools developed by multiple stakeholders. While on-call duties are inevitable, incident resolution becomes even more daunting due to the obscurity of legacy sources and the pressures of strict time constraints. To enhance the efficiency of on-call engineers (OCEs) and streamline their daily workflows, we introduced DECO -- a comprehensive framework for developing, deploying, and managing enterprise-grade chatbots tailored to improve productivity in engineering routines. This paper details the design and implementation of the DECO framework, emphasizing its innovative NL2SearchQuery functionality and a hierarchical planner. These features support efficient and customized retrieval-augmented-generation (RAG) algorithms that not only extract relevant information from diverse sources but also select the most pertinent toolkits in response to user queries. This enables the addressing of complex technical questions and provides seamless, automated access to internal resources. Additionally, DECO incorporates a robust mechanism for converting unstructured incident logs into user-friendly, structured guides, effectively bridging the documentation gap. Feedback from users underscores DECO's pivotal role in simplifying complex engineering tasks, accelerating incident resolution, and bolstering organizational productivity. Since its launch in September 2023, DECO has demonstrated its effectiveness through extensive engagement, with tens of thousands of interactions from hundreds of active users across multiple organizations within the company.

Large Language Models (LLMs), such as GPT, have revolutionized artificial intelligence by enabling nuanced understanding and generation of human-like text across a wide range of applications. However, the high computational and financial costs associated with frequent API calls to these models present a substantial bottleneck, especially for applications like customer service chatbots that handle repetitive queries. In this paper, we introduce GPT Semantic Cache, a method that leverages semantic caching of query embeddings in in-memory storage (Redis). By storing embeddings of user queries, our approach efficiently identifies semantically similar questions, allowing for the retrieval of pre-generated responses without redundant API calls to the LLM. This technique achieves a notable reduction in operational costs while significantly enhancing response times, making it a robust solution for optimizing LLM-powered applications. Our experiments demonstrate that GPT Semantic Cache reduces API calls by up to 68.8% across various query categories, with cache hit rates ranging from 61.6% to 68.8%. Additionally, the system achieves high accuracy, with positive hit rates exceeding 97%, confirming the reliability of cached responses. This technique not only reduces operational costs, but also improves response times, enhancing the efficiency of LLM-powered applications.

The Transformer architecture has revolutionized the Natural Language Processing field and is the backbone of Large Language Models (LLMs). The Transformer uses the attention mechanism that computes the pair-wise similarity between its input tokens to produce latent vectors that are able to understand the semantic meaning of the input text. One of the challenges in the Transformer architecture is the quadratic complexity of the attention mechanism that prohibits the efficient processing of long sequence lengths. One of the important works in this respect is the Perceiver class of architectures that have demonstrated excellent performance while reducing the computation complexity. In this paper, we use the PerceiverAR that was proposed for Auto-Regressive modeling as a baseline, and provide three different architectural enhancements to it with varying computation overhead tradeoffs. Inspired by the recently proposed efficient attention computation approach of Long-LoRA, we then present an equally efficient Perceiver-based architecture (termed as Long LoRA Pereceiver - LLP) that can be used as the base architecture in LLMs instead of just a fine-tuning add-on. Our results on different benchmarks indicate impressive improvements compared to recent Transformer based models. The Transformer architecture has revolutionized the field of artificial intelligence, especially in Natural Language Processing (NLP) Vaswani (2017). The recent success of Large Language models such as ChatGPT Achiam et al. (2023), Gemini Team et al. (2023), Llama Touvron et al. (2023); Dubey et al. (2024), etc. with their comprehension and reasoning capabilities, is a testament to the effectiveness of the Transformer architecture. Prior to Transformers, deep Convolutional Neural Networks (CNNs) had demonstrated amazingly well results in computer vision applications, however, their performance does not show the same effectiveness when applied to NLP.

This survey and application guide to multimodal large language models(MLLMs) explores the rapidly developing field of MLLMs, examining their architectures, applications, and impact on AI and Generative Models. Starting with foundational concepts, we delve into how MLLMs integrate various data types, including text, images, video and audio, to enable complex AI systems for cross-modal understanding and generation. It covers essential topics such as training methods, architectural components, and practical applications in various fields, from visual storytelling to enhanced accessibility. Through detailed case studies and technical analysis, the text examines prominent MLLM implementations while addressing key challenges in scalability, robustness, and cross-modal learning. Concluding with a discussion of ethical considerations, responsible AI development, and future directions, this authoritative resource provides both theoretical frameworks and practical insights. It offers a balanced perspective on the opportunities and challenges in the development and deployment of MLLMs, and is highly valuable for researchers, practitioners, and students interested in the intersection of natural language processing and computer vision.

High-Performance Computing (HPC) is crucial for performing advanced computational tasks, yet their complexity often challenges users, particularly those unfamiliar with HPC-specific commands and workflows. This paper introduces Hypothetical Command Embeddings (HyCE), a novel method that extends Retrieval-Augmented Generation (RAG) by integrating real-time, user-specific HPC data, enhancing accessibility to these systems. HyCE enriches large language models (LLM) with real-time, user-specific HPC information, addressing the limitations of fine-tuned models on such data. We evaluate HyCE using an automated RAG evaluation framework, where the LLM itself creates synthetic questions from the HPC data and serves as a judge, assessing the efficacy of the extended RAG with the evaluation metrics relevant for HPC tasks. Additionally, we tackle essential security concerns, including data privacy and command execution risks, associated with deploying LLMs in HPC environments. This solution provides a scalable and adaptable approach for HPC clusters to leverage LLMs as HPC expert, bridging the gap between users and the complex systems of HPC.