HashAttention: Semantic Sparsity for Faster Inference

Utilizing longer contexts is increasingly essential to power better AI systems. However, the cost of attending to long contexts is high due to the involved softmax computation. While the scaled dot-product attention (SDPA) exhibits token sparsity, with only a few pivotal tokens significantly contributing to attention, leveraging this sparsity effectively remains an open challenge. Previous methods either suffer from model degradation or require considerable additional resources. We propose HashAttention --a principled approach casting pivotal token identification as a recommendation problem. Given a query, HashAttention encodes keys and queries in Hamming space capturing the required semantic similarity using learned mapping functions. HashAttention efficiently identifies pivotal tokens for a given query in this Hamming space using bitwise operations, and only these pivotal tokens are used for attention computation, significantly improving overall attention efficiency. HashAttention can reduce the number of tokens used by a factor of $1/32\times$ for the Llama-3.1-8B model with LongBench, keeping average quality loss within 0.6 points, while using only 32 bits per token auxiliary memory. At $32\times$ sparsity, HashAttention is $3{-}6\times$ faster than LightLLM and $2.5{-}4.5\times$ faster than gpt-fast on Nvidia-L4 GPU.

MoE-Lightning: High-Throughput MoE Inference on Memory-constrained GPUs

Efficient deployment of large language models, particularly Mixture of Experts (MoE), on resource-constrained platforms presents significant challenges, especially in terms of computational efficiency and memory utilization. The MoE architecture, renowned for its ability to increase model capacity without a proportional increase in inference cost, greatly reduces the token generation latency compared with dense models. However, the large model size makes MoE models inaccessible to individuals without high-end GPUs. In this paper, we propose a high-throughput MoE batch inference system, that significantly outperforms past work. MoE-Lightning introduces a novel CPU-GPU-I/O pipelining schedule, CGOPipe, with paged weights to achieve high resource utilization, and a performance model, HRM, based on a Hierarchical Roofline Model we introduce to help find policies with higher throughput than existing systems. MoE-Lightning can achieve up to 10.3x higher throughput than state-of-the-art offloading-enabled LLM inference systems for Mixtral 8x7B on a single T4 GPU (16GB). When the theoretical system throughput is bounded by the GPU memory, MoE-Lightning can reach the throughput upper bound with 2-3x less CPU memory, significantly increasing resource utilization. MoE-Lightning also supports efficient batch inference for much larger MoEs (e.g., Mixtral 8x22B and DBRX) on multiple low-cost GPUs (e.g., 2-4 T4).

Drowning in Documents: Consequences of Scaling Reranker Inference

Rerankers, typically cross-encoders, are often used to re-score the documents retrieved by cheaper initial IR systems. This is because, though expensive, rerankers are assumed to be more effective. We challenge this assumption by measuring reranker performance for full retrieval, not just re-scoring first-stage retrieval. Our experiments reveal a surprising trend: the best existing rerankers provide diminishing returns when scoring progressively more documents and actually degrade quality beyond a certain limit. In fact, in this setting, rerankers can frequently assign high scores to documents with no lexical or semantic overlap with the query. We hope that our findings will spur future research to improve reranking.

Long Context RAG Performance of Large Language Models

Retrieval Augmented Generation (RAG) has emerged as a crucial technique for enhancing the accuracy of Large Language Models (LLMs) by incorporating external information. With the advent of LLMs that support increasingly longer context lengths, there is a growing interest in understanding how these models perform in RAG scenarios. Can these new long context models improve RAG performance? This paper presents a comprehensive study of the impact of increased context length on RAG performance across 20 popular open source and commercial LLMs. We ran RAG workflows while varying the total context length from 2,000 to 128,000 tokens (and 2 million tokens when possible) on three domain-specific datasets, and report key insights on the benefits and limitations of long context in RAG applications. Our findings reveal that while retrieving more documents can improve performance, only a handful of the most recent state of the art LLMs can maintain consistent accuracy at long context above 64k tokens. We also identify distinct failure modes in long context scenarios, suggesting areas for future research.

ElasticTok: Adaptive Tokenization for Image and Video

Efficient video tokenization remains a key bottleneck in learning general purpose vision models that are capable of processing long video sequences. Prevailing approaches are restricted to encoding videos to a fixed number of tokens, where too few tokens will result in overly lossy encodings, and too many tokens will result in prohibitively long sequence lengths. In this work, we introduce ElasticTok, a method that conditions on prior frames to adaptively encode a frame into a variable number of tokens. To enable this in a computationally scalable way, we propose a masking technique that drops a random number of tokens at the end of each frames's token encoding. During inference, ElasticTok can dynamically allocate tokens when needed -- more complex data can leverage more tokens, while simpler data only needs a few tokens. Our empirical evaluations on images and video demonstrate the effectiveness of our approach in efficient token usage, paving the way for future development of more powerful multimodal models, world models, and agents.

Text2SQL is Not Enough: Unifying AI and Databases with TAG

AI systems that serve natural language questions over databases promise to unlock tremendous value. Such systems would allow users to leverage the powerful reasoning and knowledge capabilities of language models (LMs) alongside the scalable computational power of data management systems. These combined capabilities would empower users to ask arbitrary natural language questions over custom data sources. However, existing methods and benchmarks insufficiently explore this setting. Text2SQL methods focus solely on natural language questions that can be expressed in relational algebra, representing a small subset of the questions real users wish to ask. Likewise, Retrieval-Augmented Generation (RAG) considers the limited subset of queries that can be answered with point lookups to one or a few data records within the database. We propose Table-Augmented Generation (TAG), a unified and general-purpose paradigm for answering natural language questions over databases. The TAG model represents a wide range of interactions between the LM and database that have been previously unexplored and creates exciting research opportunities for leveraging the world knowledge and reasoning capabilities of LMs over data. We systematically develop benchmarks to study the TAG problem and find that standard methods answer no more than 20% of queries correctly, confirming the need for further research in this area. We release code for the benchmark at https://github.com/TAG-Research/TAG-Bench.

Networks of Networks: Complexity Class Principles Applied to Compound AI Systems Design

As practitioners seek to surpass the current reliability and quality frontier of monolithic models, Compound AI Systems consisting of many language model inference calls are increasingly employed. In this work, we construct systems, which we call Networks of Networks (NoNs) organized around the distinction between generating a proposed answer and verifying its correctness, a fundamental concept in complexity theory that we show empirically extends to Language Models (LMs). We introduce a verifier-based judge NoN with K generators, an instantiation of "best-of-K" or "judge-based" compound AI systems. Through experiments on synthetic tasks such as prime factorization, and core benchmarks such as the MMLU, we demonstrate notable performance gains. For instance, in factoring products of two 3-digit primes, a simple NoN improves accuracy from 3.7\% to 36.6\%. On MMLU, a verifier-based judge construction with only 3 generators boosts accuracy over individual GPT-4-Turbo calls by 2.8\%. Our analysis reveals that these gains are most pronounced in domains where verification is notably easier than generation--a characterization which we believe subsumes many reasoning and procedural knowledge tasks, but doesn't often hold for factual and declarative knowledge-based settings. For mathematical and formal logic reasoning-based subjects of MMLU, we observe a 5-8\% or higher gain, whilst no gain on others such as geography and religion. We provide key takeaways for ML practitioners, including the importance of considering verification complexity, the impact of witness format on verifiability, and a simple test to determine the potential benefit of this NoN approach for a given problem distribution. This work aims to inform future research and practice in the design of compound AI systems.

LOTUS: Enabling Semantic Queries with LLMs Over Tables of Unstructured and Structured Data

The semantic capabilities of language models (LMs) have the potential to enable rich analytics and reasoning over vast knowledge corpora. Unfortunately, existing systems lack high-level abstractions to perform semantic queries at scale. We introduce semantic operators, a declarative programming interface that extends the relational model with composable AI-based operations for semantic queries over datasets (e.g., sorting or aggregating records using natural language criteria). Each operator can be implemented and optimized in multiple ways, opening a rich space for execution plans similar to relational operators. We implement our operators and several optimizations for them in LOTUS, an open-source query engine with a Pandas-like API. We demonstrate LOTUS' effectiveness across a series of real applications, including fact-checking, extreme multi-label classification, and search. We find that LOTUS' programming model is highly expressive, capturing state-of-the-art query pipelines with low development overhead. Specifically, on the FEVER dataset, LOTUS' programs can reproduce FacTool, a recent state-of-the-art fact-checking pipeline, in few lines of code, and implement a new pipeline that improves accuracy by $9.5\%$, while offering $7-34\times$ lower execution time. In the extreme multi-label classification task on the BioDEX dataset, LOTUS reproduces state-of-the art result quality with its join operator, while providing an efficient algorithm that runs $800\times$ faster than a naive join. In the search and ranking application, LOTUS allows a simple composition of operators to achieve $5.9 - 49.4\%$ higher nDCG@10 than the vanilla retriever and re-ranker, while also providing query efficiency, with $1.67 - 10\times$ lower execution time than LM-based ranking methods used by prior works. LOTUS is publicly available at https://github.com/stanford-futuredata/lotus.

Optimizing Instructions and Demonstrations for Multi-Stage Language Model Programs

Language Model Programs, i.e. sophisticated pipelines of modular language model (LM) calls, are increasingly advancing NLP tasks, but they require crafting prompts that are jointly effective for all modules. We study prompt optimization for LM programs, i.e. how to update these prompts to maximize a downstream metric without access to module-level labels or gradients. To make this tractable, we factorize our problem into optimizing the free-form instructions and few-shot demonstrations of every module and introduce several strategies to craft task-grounded instructions and navigate credit assignment across modules. Our strategies include (i) program- and data-aware techniques for proposing effective instructions, (ii) a stochastic mini-batch evaluation function for learning a surrogate model of our objective, and (iii) a meta-optimization procedure in which we refine how LMs construct proposals over time. Using these insights we develop MIPRO, a novel optimizer that outperforms baselines on five of six diverse LM programs using a best-in-class open-source model (Llama-3-8B), by as high as 12.9% accuracy. We will release our new optimizers and benchmark in DSPy at https://github.com/stanfordnlp/dspy

Generating Probabilistic Scenario Programs from Natural Language

For cyber-physical systems (CPS), including robotics and autonomous vehicles, mass deployment has been hindered by fatal errors that occur when operating in rare events. To replicate rare events such as vehicle crashes, many companies have created logging systems and employed crash reconstruction experts to meticulously recreate these valuable events in simulation. However, in these methods, "what if" questions are not easily formulated and answered. We present ScenarioNL, an AI System for creating scenario programs from natural language. Specifically, we generate these programs from police crash reports. Reports normally contain uncertainty about the exact details of the incidents which we represent through a Probabilistic Programming Language (PPL), Scenic. By using Scenic, we can clearly and concisely represent uncertainty and variation over CPS behaviors, properties, and interactions. We demonstrate how commonplace prompting techniques with the best Large Language Models (LLM) are incapable of reasoning about probabilistic scenario programs and generating code for low-resource languages such as Scenic. Our system is comprised of several LLMs chained together with several kinds of prompting strategies, a compiler, and a simulator. We evaluate our system on publicly available autonomous vehicle crash reports in California from the last five years and share insights into how we generate code that is both semantically meaningful and syntactically correct.