arXiv subjects

Aditya Agrawal

Publications and source records attributed to Aditya Agrawal.

16 recordsLinked to original sources

arXiv2026

WARP: Wasserstein-Aligned RAG for Population Opinions

RAG systems are increasingly used to summarize what large collections of documents say. A user asks "What do people think about X?" and receives an answer that reads as consensus. But standard top-k retrieval ranks documents by query similarity, not by how faithfully they represent the population, so minority views quietly disappear. Existing fixes fall short. Diversity re-rankers like MMR and DPP spread retrieved documents apart, but with no target distribution to aim for. Calibration methods based on KL or JS divergence do target one, yet treat opinion bins as unordered: confusing strong positive with strong negative costs no more than an adjacent-bin miss. We introduce WARP, a family of post-retrieval algorithms that calibrate retrieved evidence to the population's opinion distribution. WARP first recovers underrepresented opinions that cosine ranking may bury, then uses Wasserstein-1 distance to select documents whose sentiment-intensity distribution matches the population target, capturing the ordinal structure ignored by KL and JS divergence. We develop three variants for dense, sparse, and variable candidate pools, trading off calibration quality and speed. Across three review domains spanning 35K documents, 156 queries, and 26 entities, WARP's domain-matched variants reduce distributional error by at least 43% with sub-second latency. These gains carry through to generation: a five-judge LLM panel prefers WARP-generated answers in 86% of decided comparisons at k <= 5.

cs.IR↗

arXiv2026

Act2See: Emergent Active Visual Perception for Video Reasoning

Vision-Language Models (VLMs) typically rely on static initial frames for video reasoning, restricting their ability to incorporate essential dynamic information as the reasoning process evolves. Existing methods that augment Chain-of-Thought (CoT) with additional frame information often exhibit suboptimal CoT quality and lack the crucial ability to synthesize visual information for hypothetical or counterfactual scenarios. We introduce Act-to-See (Act2See), a novel framework that enables active visual perception by empowering VLMs to actively interleave video frames within text CoTs. Act2See is developed via Supervised Fine-Tuning (SFT) on a high-quality dataset of reasoning traces generated by a frontier VLM. These traces integrate active calls to either retrieve existing frames or generate new ones, and are rigorously verified against human-annotated CoTs to ensure quality. This approach cultivates an emergent capability: at inference time, the model actively determines when to search for or synthesize the necessary visual evidence. Act2See establishes new state-of-the-art results on challenging benchmarks, including VideoEspresso and ViTIB, and outperforms comparable or larger models on Video-MME, EgoNormia, and VCR-Bench, demonstrating an advancement in enabling VLMs with active visual perception for video reasoning.

cs.CV↗

arXiv2026

Video Active Perception: Effective Inference-Time Long-Form Video Understanding with Vision-Language Models

Large vision-language models (VLMs) have advanced multimodal tasks such as video question answering (QA). However, VLMs face the challenge of selecting frames effectively and efficiently, as standard uniform sampling is expensive and performance may plateau. Inspired by active perception theory, which posits that models gain information by acquiring data that differs from their expectations, we introduce Video Active Perception (VAP), a training-free method to enhance long-form video QA using VLMs. Our approach treats keyframe selection as data acquisition in active perception and leverages a lightweight text-conditioned video generation model to represent prior world knowledge. Empirically, VAP achieves state-of-the-art zero-shot results on long-form or reasoning video QA datasets such as EgoSchema, NExT-QA, ActivityNet-QA, IntentQA, and CLEVRER, achieving an increase of up to 5.6 x frame efficiency by frames per question over standard GPT-4o, Gemini 1.5 Pro, and LLaVA-OV. Moreover, VAP shows stronger reasoning abilities than previous methods and effectively selects keyframes relevant to questions. These findings highlight the potential of leveraging active perception to improve the frame effectiveness and efficiency of long-form video QA.

cs.CV↗

arXiv2026

Nemotron 3 Super: Open, Efficient Mixture-of-Experts Hybrid Mamba-Transformer Model for Agentic Reasoning

We describe the pre-training, post-training, and quantization of Nemotron 3 Super, a 120 billion (active 12 billion) parameter hybrid Mamba-Attention Mixture-of-Experts model. Nemotron 3 Super is the first model in the Nemotron 3 family to 1) be pre-trained in NVFP4, 2) leverage LatentMoE, a new Mixture-of-Experts architecture that optimizes for both accuracy per FLOP and accuracy per parameter, and 3) include MTP layers for inference acceleration through native speculative decoding. We pre-trained Nemotron 3 Super on 25 trillion tokens followed by post-training using supervised fine tuning (SFT) and reinforcement learning (RL). The final model supports up to 1M context length and achieves comparable accuracy on common benchmarks, while also achieving up to 2.2x and 7.5x higher inference throughput compared to GPT-OSS-120B and Qwen3.5-122B, respectively. Nemotron 3 Super datasets, along with the base, post-trained, and quantized checkpoints, are open-sourced on HuggingFace.

cs.LG↗

arXiv2026

Retrieval-Augmented Generation Must Move Beyond Factual Grounding to Represent Diverse Opinions

This position paper argues that Retrieval-Augmented Generation (RAG) systems exhibit a factual bias-optimizing for epistemic uncertainty reduction while ignoring the aleatoric uncertainty inherent in opinion-rich content. This misalignment demands a paradigm shift in RAG system design. A survey of 34 major RAG benchmarks reveals that only one addresses opinion synthesis, confirming that the bias is structural and embedded in datasets, retrieval-generation objectives, and evaluation metrics alike. Beyond technical limitations, this bias poses risks to transparent and accountable AI. Namely, echo chamber effects that amplify dominant viewpoints, which can lead to opinion manipulation and under-representation of minority voices. We formalize the problem through the lens of uncertainty quantification, showing that factual queries should minimize posterior entropy while opinion queries must preserve it. We derive a unified objective over coverage, fidelity, and fairness using the Wasserstein distance. As an existence proof, we present Opinion-Aware RAG (O-RAG), an architecture featuring LLM-based opinion extraction and entity-linked opinion metadata. We evaluate it across two domains -- e-commerce seller forums and public hotel reviews. Experiments demonstrate 18-48% reduction in Wasserstein distance to corpus-level sentiment distributions, +26.8% sentiment diversity, and +42.7% entity match rate. Human evaluators preferred opinion-enriched generation 79.2% of the time. We propose a research agenda and argue that as RAG systems increasingly mediate access to information, their ability to represent diverse perspectives is of the essence.

cs.AI↗

arXiv2026

Quad Length Codes for Lossless Compression of e4m3

Training and serving Large Language Models (LLMs) relies heavily on parallelization and collective operations, which are frequently bottlenecked by network bandwidth. Lossless compression using e.g., Huffman codes can alleviate the issue, however, Huffman codes suffer from slow, bit-sequential decoding and high hardware complexity due to deep tree traversals. Universal codes e.g., Exponential-Golomb codes are faster to decode but do not exploit the symbol frequency distributions. To address these limitations, this paper introduces Quad Length Codes, a hybrid approach designed to balance compression efficiency with decoding speed. The coding scheme uses 3 prefix bits to divide the 256 symbols into 8 areas. Each area has a different code length and encodes a different number of symbols. The scheme uses a Look Up Table with 256 entries, significantly simplifying the hardware implementation compared to Huffman trees. The coding scheme can be adapted for different distributions. For the e4m3 data type, the scheme achieves a compressibility of 13.9% in comparison to 15.9% achieved by Huffman codes, but it significantly speeds up the decoding and simplifies the hardware complexity.

cs.LG↗

arXiv2026

Single-Stage Huffman Encoder for ML Compression

Training and serving Large Language Models (LLMs) require partitioning data across multiple accelerators, where collective operations are frequently bottlenecked by network bandwidth. Lossless compression using Huffman codes is an effective way to alleviate the issue, however, its three-stage design requiring on-the-fly frequency analysis, codebook generation and transmission of codebook along with data introduces computational, latency and data overheads which are prohibitive for latency-sensitive scenarios such as die-to-die communication. This paper proposes a single-stage Huffman encoder that eliminates these overheads by using fixed codebooks derived from the average probability distribution of previous data batches. Through our analysis of the Gemma 2B model, we demonstrate that tensors exhibit high statistical similarity across layers and shards. Using this approach we achieve compression within 0.5% of per-shard Huffman coding and within 1% of the ideal Shannon compressibility, enabling efficient on-the-fly compression.

cs.LG↗

arXiv2025

PoseBench3D: A Cross-Dataset Analysis Framework for 3D Human Pose Estimation via Pose Lifting Networks

Reliable three-dimensional human pose estimation (3D HPE) remains challenging due to the differences in viewpoints, environments, and camera conventions among datasets. As a result, methods that achieve near-optimal in-dataset accuracy often degrade on unseen datasets. In practice, however, systems must adapt to diverse viewpoints, environments, and camera setups--conditions that differ significantly from those encountered during training, which is often the case in real-world scenarios. Measuring cross-dataset performance is a vital process, but extremely labor-intensive when done manually for human pose estimation. To address these challenges, we automate this evaluation using PoseBench3D, a standardized testing framework that enables consistent and fair cross-dataset comparisons on previously unseen data. PoseBench3D streamlines testing across four widely used 3D HPE datasets via a single, configurable interface. Using this framework, we re-evaluate 18 methods and report over 100 cross-dataset results under Protocol 1: MPJPE and Protocol 2: PA-MPJPE, revealing systematic generalization gaps and the impact of common preprocessing and dataset setup choices. The PoseBench3D code is found at: https://github.com/bryanjvela/PoseBench3D

cs.CV↗

arXiv2024

Latent Representation Matters: Human-like Sketches in One-shot Drawing Tasks

Humans can effortlessly draw new categories from a single exemplar, a feat that has long posed a challenge for generative models. However, this gap has started to close with recent advances in diffusion models. This one-shot drawing task requires powerful inductive biases that have not been systematically investigated. Here, we study how different inductive biases shape the latent space of Latent Diffusion Models (LDMs). Along with standard LDM regularizers (KL and vector quantization), we explore supervised regularizations (including classification and prototype-based representation) and contrastive inductive biases (using SimCLR and redundancy reduction objectives). We demonstrate that LDMs with redundancy reduction and prototype-based regularizations produce near-human-like drawings (regarding both samples' recognizability and originality) -- better mimicking human perception (as evaluated psychophysically). Overall, our results suggest that the gap between humans and machines in one-shot drawings is almost closed.

cs.CV↗

arXiv2024

eXmY: A Data Type and Technique for Arbitrary Bit Precision Quantization

eXmY is a novel data type for quantization of ML models. It supports both arbitrary bit widths and arbitrary integer and floating point formats. For example, it seamlessly supports 3, 5, 6, 7, 9 bit formats. For a specific bit width, say 7, it defines all possible formats e.g. e0m6, e1m5, e2m4, e3m3, e4m2, e5m1 and e6m0. For non-power of two bit widths e.g. 5, 6, 7, we created a novel encoding and decoding scheme which achieves perfect compression, byte addressability and is amenable to sharding and vector processing. We implemented libraries for emulation, encoding and decoding tensors and checkpoints in C++, TensorFlow, JAX and PAX. For optimal performance, the codecs use SIMD instructions on CPUs and vector instructions on TPUs and GPUs. eXmY is also a technique and exploits the statistical distribution of exponents in tensors. It can be used to quantize weights, static and dynamic activations, gradients, master weights and optimizer state. It can reduce memory (CPU DRAM and accelerator HBM), network and disk storage and transfers. It can increase multi tenancy and accelerate compute. eXmY has been deployed in production for almost 2 years.

cs.LG↗

arXiv2024

Large Content And Behavior Models To Understand, Simulate, And Optimize Content And Behavior

Shannon and Weaver's seminal information theory divides communication into three levels: technical, semantic, and effectiveness. While the technical level deals with the accurate reconstruction of transmitted symbols, the semantic and effectiveness levels deal with the inferred meaning and its effect on the receiver. Large Language Models (LLMs), with their wide generalizability, make some progress towards the second level. However, LLMs and other communication models are not conventionally designed for predicting and optimizing communication for desired receiver behaviors and intents. As a result, the effectiveness level remains largely untouched by modern communication systems. In this paper, we introduce the receivers' "behavior tokens," such as shares, likes, clicks, purchases, and retweets, in the LLM's training corpora to optimize content for the receivers and predict their behaviors. Other than showing similar performance to LLMs on content understanding tasks, our trained models show generalization capabilities on the behavior dimension for behavior simulation, content simulation, behavior understanding, and behavior domain adaptation. We show results on all these capabilities using a wide range of tasks on three corpora. We call these models Large Content and Behavior Models (LCBMs). Further, to spur more research on LCBMs, we release our new Content Behavior Corpus (CBC), a repository containing communicator, message, and corresponding receiver behavior (https://behavior-in-the-wild.github.io/LCBM).

cs.CL↗

arXiv2023

COVID-19 Vaccine Misinformation in Middle Income Countries

This paper introduces a multilingual dataset of COVID-19 vaccine misinformation, consisting of annotated tweets from three middle-income countries: Brazil, Indonesia, and Nigeria. The expertly curated dataset includes annotations for 5,952 tweets, assessing their relevance to COVID-19 vaccines, presence of misinformation, and the themes of the misinformation. To address challenges posed by domain specificity, the low-resource setting, and data imbalance, we adopt two approaches for developing COVID-19 vaccine misinformation detection models: domain-specific pre-training and text augmentation using a large language model. Our best misinformation detection models demonstrate improvements ranging from 2.7 to 15.9 percentage points in macro F1-score compared to the baseline models. Additionally, we apply our misinformation detection models in a large-scale study of 19 million unlabeled tweets from the three countries between 2020 and 2022, showcasing the practical application of our dataset and models for detecting and analyzing vaccine misinformation in multiple countries and languages. Our analysis indicates that percentage changes in the number of new COVID-19 cases are positively associated with COVID-19 vaccine misinformation rates in a staggered manner for Brazil and Indonesia, and there are significant positive associations between the misinformation rates across the three countries.

cs.CL↗

arXiv2022

Neural Feature-Adaptation for Symbolic Predictions Using Pre-Training and Semantic Loss

We are interested in neurosymbolic systems consisting of a high-level symbolic layer for explainable prediction in terms of human-intelligible concepts; and a low-level neural layer for extracting symbols required to generate the symbolic explanation. Real data is often imperfect meaning that even if the symbolic theory remains unchanged, we may still need to address the problem of mapping raw data to high-level symbols, each time there is a change in the data acquisition environment or equipment. Manual (re-)annotation of the raw data each time this happens is laborious and expensive; and automated labelling methods are often imperfect, especially for complex problems. NEUROLOG proposed the use of a semantic loss function that allows an existing feature-based symbolic model to guide the extraction of feature-values from raw data, using `abduction'. However, the experiments demonstrating the use of semantic loss through abduction appear to rely heavily on a domain-specific pre-processing step that enables a prior delineation of feature locations in the raw data. We examine the use of semantic loss in domains where such pre-processing is not possible, or is not obvious. We show that without any prior information about the features, the NEUROLOG approach can continue to predict accurately even with substantially incorrect feature predictions. We show also that prior information about the features in the form of even imperfect pre-training can help correct this situation. These findings are replicated on the original problem considered by NEUROLOG, without the use of feature-delineation. This suggests that symbolic explanations constructed for data in a domain could be re-used in a related domain, by `feature-adaptation' of pre-trained neural extractors using the semantic loss function constrained by abductive feedback.

cs.AI↗

arXiv2018

What Your DRAM Power Models Are Not Telling You: Lessons from a Detailed Experimental Study

Main memory (DRAM) consumes as much as half of the total system power in a computer today, resulting in a growing need to develop new DRAM architectures and systems that consume less power. Researchers have long relied on DRAM power models that are based off of standardized current measurements provided by vendors, called IDD values. Unfortunately, we find that these models are highly inaccurate, and do not reflect the actual power consumed by real DRAM devices. We perform the first comprehensive experimental characterization of the power consumed by modern real-world DRAM modules. Our extensive characterization of 50 DDR3L DRAM modules from three major vendors yields four key new observations about DRAM power consumption: (1) across all IDD values that we measure, the current consumed by real DRAM modules varies significantly from the current specified by the vendors; (2) DRAM power consumption strongly depends on the data value that is read or written; (3) there is significant structural variation, where the same banks and rows across multiple DRAM modules from the same model consume more power than other banks or rows; and (4) over successive process technology generations, DRAM power consumption has not decreased by as much as vendor specifications have indicated. Based on our detailed analysis and characterization data, we develop the Variation-Aware model of Memory Power Informed by Real Experiments (VAMPIRE). We show that VAMPIRE has a mean absolute percentage error of only 6.8% compared to actual measured DRAM power. VAMPIRE enables a wide range of studies that were not possible using prior DRAM power models. As an example, we use VAMPIRE to evaluate a new power-aware data encoding mechanism, which can reduce DRAM energy consumption by an average of 12.2%. We plan to open-source both VAMPIRE and our extensive raw data collected during our experimental characterization.

cs.AR↗

arXiv2018

Voltron: Understanding and Exploiting the Voltage-Latency-Reliability Trade-Offs in Modern DRAM Chips to Improve Energy Efficiency

This paper summarizes our work on experimental characterization and analysis of reduced-voltage operation in modern DRAM chips, which was published in SIGMETRICS 2017, and examines the work's significance and future potential. We take a comprehensive approach to understanding and exploiting the latency and reliability characteristics of modern DRAM when the DRAM supply voltage is lowered below the nominal voltage level specified by DRAM standards. We perform an experimental study of 124 real DDR3L (low-voltage) DRAM chips manufactured recently by three major DRAM vendors. We find that reducing the supply voltage below a certain point introduces bit errors in the data, and we comprehensively characterize the behavior of these errors. We discover that these errors can be avoided by increasing the latency of three major DRAM operations (activation, restoration, and precharge). We perform detailed DRAM circuit simulations to validate and explain our experimental findings. We also characterize the various relationships between reduced supply voltage and error locations, stored data patterns, DRAM temperature, and data retention. Based on our observations, we propose a new DRAM energy reduction mechanism, called Voltron. The key idea of Voltron is to use a performance model to determine by how much we can reduce the supply voltage without introducing errors and without exceeding a user-specified threshold for performance loss. Our evaluations show that Voltron reduces the average DRAM and system energy consumption by 10.5% and 7.3%, respectively, while limiting the average system performance loss to only 1.8%, for a variety of memory-intensive quad-core workloads. We also show that Voltron significantly outperforms prior dynamic voltage and frequency scaling mechanisms for DRAM.

cs.AR↗

arXiv2017

Understanding Reduced-Voltage Operation in Modern DRAM Chips: Characterization, Analysis, and Mechanisms

The energy consumption of DRAM is a critical concern in modern computing systems. Improvements in manufacturing process technology have allowed DRAM vendors to lower the DRAM supply voltage conservatively, which reduces some of the DRAM energy consumption. We would like to reduce the DRAM supply voltage more aggressively, to further reduce energy. Aggressive supply voltage reduction requires a thorough understanding of the effect voltage scaling has on DRAM access latency and DRAM reliability. In this paper, we take a comprehensive approach to understanding and exploiting the latency and reliability characteristics of modern DRAM when the supply voltage is lowered below the nominal voltage level specified by DRAM standards. Using an FPGA-based testing platform, we perform an experimental study of 124 real DDR3L (low-voltage) DRAM chips manufactured recently by three major DRAM vendors. We find that reducing the supply voltage below a certain point introduces bit errors in the data, and we comprehensively characterize the behavior of these errors. We discover that these errors can be avoided by increasing the latency of three major DRAM operations (activation, restoration, and precharge). We perform detailed DRAM circuit simulations to validate and explain our experimental findings. We also characterize the various relationships between reduced supply voltage and error locations, stored data patterns, DRAM temperature, and data retention. Based on our observations, we propose a new DRAM energy reduction mechanism, called Voltron. The key idea of Voltron is to use a performance model to determine by how much we can reduce the supply voltage without introducing errors and without exceeding a user-specified threshold for performance loss. Voltron reduces the average system energy by 7.3% while limiting the average system performance loss to only 1.8%, for a variety of workloads.

cs.AR↗