arXiv subjects

Ankur Srivastava

Publications and source records attributed to Ankur Srivastava.

16 recordsLinked to original sources

arXiv2026

WP-MIP: An Artificial Intelligence, Hybrid, and Physically Based Model Intercomparison Project for Weather Prediction

Rapid progress in the field of machine-learning for weather prediction has led to the emergence of algorithms whose forecasting skill can exceed that of traditional physically based models. This development represents an opportunity to improve the quality of forecasting services provided by operational centers, particularly given the speed at which machine-learning based models generate predictions. Despite the clear promise of these systems, questions remain about the ability of the current generation of machine-learning models to generate physically consistent predictions of the full suite of required forecast fields under all conditions. Answering these questions will require careful comparisons between the well-understood physically based models, current state-of-the-art machine-learning models, and the hybrid models that combine elements of these two archetypes. The Weather Prediction Model Intercomparison Project (WP-MIP) is a World Meteorological Organization-supported initiative whose initial goal is to create a centralized database of physically based, machine-learning, and hybrid model forecasts to enable a distributed assessment and evaluation effort. The first instance of WP-MIP focuses on global deterministic predictions using both center-specific and common initializations to facilitate sensitivity studies. Forecasts contributed by institutions across six continents will be used to develop AI-ready verification techniques that highlight the strengths and weaknesses of each class of prediction system, with the goal of establishing best-practice guidance to model developers and national weather centers. The broad engagement of the operational and forecast-evaluation communities in WP-MIP will ensure that the project results are highly relevant to the development and deployment of next-generation weather prediction systems.

physics.ao-ph↗

arXiv2026

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

We introduce Nemotron 3 Ultra, a 550 billion total and 55 billion active parameter Mixture-of-Experts Hybrid Mamba-Attention language model. We pre-trained Nemotron 3 Ultra on 20 trillion text tokens, then extended the context length to 1M tokens, and post-trained using Supervised Fine Tuning (SFT), Reinforcement Learning (RL), and Multi-teacher On-Policy Distillation (MOPD). Nemotron 3 Ultra is our most capable model yet, employing multiple key technologies - LatentMoE, Multi Token Prediction (MTP), NVFP4 pre-training, multi-environment RLVR, MOPD, and reasoning budget control. Nemotron 3 Ultra achieves up to ~6x higher inference throughput as compared to state-of-the-art publicly available LLMs while attaining on-par accuracy. The state-of-the-art accuracy, high inference throughput, and 1M token context length make Nemotron 3 Ultra ideal for long-running autonomous agentic tasks. We open-source the base, post-trained, and quantized checkpoints, along with the training data and recipe on HuggingFace.

cs.CL↗

arXiv2026

Experimental and computational diffusion analysis in Ni-X binary and Ni-Al-X (X = Cr, Mo, Ta, W, Re) ternary systems

An extensive diffusion analysis is presented for binary Ni-X and ternary Ni-Al-X (X = Cr, Mo, Ta, W, Re) systems, which play a crucial role in microstructural evolution and phase stability in Ni-Al-based superalloys. Specifically, we highlight changes in the diffusion coefficients of X in the presence of Al and compare diffusional interactions across systems considered. First-principles calculations, combined with activation energies derived from temperature-dependent experiments, reveal consistent trends in Ni-X systems, with variations in activation energies largely attributed to differences in migration energies. In ternary systems, diffusion coefficients estimated from intersecting diffusion profiles show that the main interdiffusion coefficient of X is comparable to its binary counterpart, with similar activation energies. However, cross-diffusion coefficients are shown to significantly influence fluxes, either enhancing or reducing diffusion lengths depending on the relative directions of diffusing elements. For Ni-Al-Re, a single-profile method is employed to overcome uncertainties in estimating composition gradients at the near-end-member intersecting composition. The diffusion coefficients obtained correlate well with the nature of diffusion paths when represented on Gibbs triangles. To extend these findings, a physics-informed neural network (PINN) optimization method is applied to extract composition-dependent diffusion coefficients across the full composition range. The analysis demonstrates the necessity of incorporating experimentally estimated diffusion coefficients as equality constraints, without which optimization reliability is compromised. Overall, the results establish a robust framework for diffusion studies in Ni-Al-X systems, highlighting the critical role of cross-diffusion effects and constraint-enhanced numerical methods.

cond-mat.mtrl-sci↗

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↗

arXiv2025

SymRTLO: Enhancing RTL Code Optimization with LLMs and Neuron-Inspired Symbolic Reasoning

Optimizing Register Transfer Level (RTL) code is crucial for improving the power, performance, and area (PPA) of digital circuits in the early stages of synthesis. Manual rewriting, guided by synthesis feedback, can yield high-quality results but is time-consuming and error-prone. Most existing compiler-based approaches have difficulty handling complex design constraints. Large Language Model (LLM)-based methods have emerged as a promising alternative to address these challenges. However, LLM-based approaches often face difficulties in ensuring alignment between the generated code and the provided prompts. This paper presents SymRTLO, a novel neuron-symbolic RTL optimization framework that seamlessly integrates LLM-based code rewriting with symbolic reasoning techniques. Our method incorporates a retrieval-augmented generation (RAG) system of optimization rules and Abstract Syntax Tree (AST)-based templates, enabling LLM-based rewriting that maintains syntactic correctness while minimizing undesired circuit behaviors. A symbolic module is proposed for analyzing and optimizing finite state machine (FSM) logic, allowing fine-grained state merging and partial specification handling beyond the scope of pattern-based compilers. Furthermore, a fast verification pipeline, combining formal equivalence checks with test-driven validation, further reduces the complexity of verification. Experiments on the RTL-Rewriter benchmark with Synopsys Design Compiler and Yosys show that SymRTLO improves power, performance, and area (PPA) by up to 43.9%, 62.5%, and 51.1%, respectively, compared to the state-of-the-art methods.

cs.AR↗

arXiv2025

TroLL: Exploiting Structural Similarities between Logic Locking and Hardware Trojans

Logic locking and hardware Trojans are two fields in hardware security that have been mostly developed independently from each other. In this paper, we identify the relationship between these two fields. We find that a common structure that exists in many logic locking techniques has desirable properties of hardware Trojans (HWT). We then construct a novel type of HWT, called Trojans based on Logic Locking (TroLL), in a way that can evade state-of-the-art ATPG-based HWT detection techniques. In an effort to detect TroLL, we propose customization of existing state-of-the-art ATPG-based HWT detection approaches as well as adapting the SAT-based attacks on logic locking to HWT detection. In our experiments, we use random sampling as reference. It is shown that the customized ATPG-based approaches are the best performing but only offer limited improvement over random sampling. Moreover, their efficacy also diminishes as TroLL's triggers become longer (i.e. have more bits specified). We thereby highlight the need to find a scalable HWT detection approach for TroLL.

cs.CR↗

arXiv2025

Red Teaming Methodology for Design Obfuscation

The main goal of design obfuscation schemes is to protect sensitive design details from untrusted parties in the VLSI supply chain, including but not limited to off-shore foundries and untrusted end users. In this work, we provide a systematic red teaming approach to evaluate the security of design obfuscation approaches. Specifically, we propose security metrics and evaluation methodology for the scenarios where the adversary does not have access to a working chip. A case study on the RIPPER tool developed by the University of Florida indicates that more information is leaked about the structure of the original design than commonly considered.

cs.CR↗

arXiv2025

Hardware-Enabled Mechanisms for Verifying Responsible AI Development

Advancements in AI capabilities, driven in large part by scaling up computing resources used for AI training, have created opportunities to address major global challenges but also pose risks of misuse. Hardware-enabled mechanisms (HEMs) can support responsible AI development by enabling verifiable reporting of key properties of AI training activities such as quantity of compute used, training cluster configuration or location, as well as policy enforcement. Such tools can promote transparency and improve security, while addressing privacy and intellectual property concerns. Based on insights from an interdisciplinary workshop, we identify open questions regarding potential implementation approaches, emphasizing the need for further research to ensure robust, scalable solutions.

cs.CR↗

arXiv2023

Role of modified cloud microphysics parameterization in coupled climate model for studying ISM rainfall: small-scale cloud model and climate model work better together

An unresolved problem of present generation coupled climate models is the realistic distribution of rainfall over Indian monsoon region, which is also related to the persistent dry bias over Indian land mass. Therefore, quantitative prediction of the intensity of rainfall events has remained a challenge for the state-of-the-art global coupled models. Guided by the observation, it is hypothesized that insufficient growth of cloud droplets and processes responsible for the cloud to rain water conversion are key components to distinguish between shallow to convective clouds. The new diffusional growth rates and relative dispersion based autoconversion from the Eulerian-Lagrangian particleby-particle based small-scale model provide a pathway to revisit the parameterizations in climate models for monsoon clouds. The realistic information of cloud drop size distribution is incorporated in the microphysical parameterization scheme of climate model. Two sensitivity simulations are conducted using coupled forecast system (CFSv2) model. When our physically based small-scale derived modified parameterization is used, a coupled climate model simulates the probability distribution (PDF) of rainfall and accompanying specific humidity, liquid water content, and outgoing long-wave radiation (OLR) with increasing accuracy. The improved simulation of rainfall PDF appears to have been aided by much improved simulation of OLR and resulted better simulation of the ISM rainfall.

physics.ao-ph↗

arXiv2022

DynaMarks: Defending Against Deep Learning Model Extraction Using Dynamic Watermarking

The functionality of a deep learning (DL) model can be stolen via model extraction where an attacker obtains a surrogate model by utilizing the responses from a prediction API of the original model. In this work, we propose a novel watermarking technique called DynaMarks to protect the intellectual property (IP) of DL models against such model extraction attacks in a black-box setting. Unlike existing approaches, DynaMarks does not alter the training process of the original model but rather embeds watermark into a surrogate model by dynamically changing the output responses from the original model prediction API based on certain secret parameters at inference runtime. The experimental outcomes on Fashion MNIST, CIFAR-10, and ImageNet datasets demonstrate the efficacy of DynaMarks scheme to watermark surrogate models while preserving the accuracies of the original models deployed in edge devices. In addition, we also perform experiments to evaluate the robustness of DynaMarks against various watermark removal strategies, thus allowing a DL model owner to reliably prove model ownership.

cs.CR↗

arXiv2021

Robust and Attack Resilient Logic Locking with a High Application-Level Impact

Logic locking is a hardware security technique to intellectual property (IP) against security threats in the IC supply chain, especially untrusted fabs. Such techniques incorporate additional locking circuitry within an IC that induces incorrect functionality when an incorrect key is provided. The amount of error induced is known as the effectiveness of the locking technique. "SAT attacks" provide a strong mathematical formulation to find the correct key of locked circuits. In order to achieve high SAT resilience(i.e. complexity of SAT attacks), many conventional logic locking schemes fail to inject sufficient error into the circuit. For example, in the case of SARLock and Anti-SAT, there are usually very few (or only one) input minterms that cause any error at the circuit output. The state-of-the-art stripped functionality logic locking (SFLL) technique introduced a trade-off between SAT resilience and effectiveness. In this work, we prove that such a trade-off is universal in logic locking. In order to attain high effectiveness of locking without compromising SAT resilience, we propose a novel logic locking scheme, called Strong Anti-SAT (SAS). In addition to SAT attacks, removal-based attacks are also popular against logic locking. Based on SAS, we propose Robust SAS (RSAS) which is resilient to removal attacks and maintains the same SAT resilience and as effectiveness as SAS. SAS and RSAS have the following significant improvements over existing techniques. (1) SAT resilience of SAS and RSAS against SAT attack is not compromised by increase in effectiveness. (2) In contrast to prior work focusing solely on the circuit-level locking impact, we integrate SAS-locked modules into a processor and show that SAS has a high application-level impact. (3) Our experiments show that SAS and RSAS exhibit better SAT resilience than SFLL and have similar effectiveness.

cs.CR↗

arXiv2020

Benchmarking at the Frontier of Hardware Security: Lessons from Logic Locking

Integrated circuits (ICs) are the foundation of all computing systems. They comprise high-value hardware intellectual property (IP) that are at risk of piracy, reverse-engineering, and modifications while making their way through the geographically-distributed IC supply chain. On the frontier of hardware security are various design-for-trust techniques that claim to protect designs from untrusted entities across the design flow. Logic locking is one technique that promises protection from the gamut of threats in IC manufacturing. In this work, we perform a critical review of logic locking techniques in the literature, and expose several shortcomings. Taking inspiration from other cybersecurity competitions, we devise a community-led benchmarking exercise to address the evaluation deficiencies. In reflecting on this process, we shed new light on deficiencies in evaluation of logic locking and reveal important future directions. The lessons learned can guide future endeavors in other areas of hardware security.

cs.CR↗

arXiv2018

In-situ Stochastic Training of MTJ Crossbar based Neural Networks

Owing to high device density, scalability and non-volatility, Magnetic Tunnel Junction-based crossbars have garnered significant interest for implementing the weights of an artificial neural network. The existence of only two stable states in MTJs implies a high overhead of obtaining optimal binary weights in software. We illustrate that the inherent parallelism in the crossbar structure makes it highly appropriate for in-situ training, wherein the network is taught directly on the hardware. It leads to significantly smaller training overhead as the training time is independent of the size of the network, while also circumventing the effects of alternate current paths in the crossbar and accounting for manufacturing variations in the device. We show how the stochastic switching characteristics of MTJs can be leveraged to perform probabilistic weight updates using the gradient descent algorithm. We describe how the update operations can be performed on crossbars both with and without access transistors and perform simulations on them to demonstrate the effectiveness of our techniques. The results reveal that stochastically trained MTJ-crossbar NNs achieve a classification accuracy nearly same as that of real-valued-weight networks trained in software and exhibit immunity to device variations.

cs.NE↗

arXiv2018

Data Driven Optimizations for MTJ based Stochastic Computing

Stochastic computing, a form of computation with probabilities, presents an alternative to conventional arithmetic units. Magnetic Tunnel Junctions (MTJs), which exhibit probabilistic switching, have been explored as Stochastic Number Generators (SNGs). We provide a perspective of the energy requirements of such an application and design an energy-efficient and data-sensitive MTJ-based SNG. We discuss its benefits when used for stochastic computations, illustrating with the help of a multiplier circuit, in terms of energy savings when compared to computing with the baseline MTJ-SNG.

cs.ET↗

arXiv2017

Neural Trojans

While neural networks demonstrate stronger capabilities in pattern recognition nowadays, they are also becoming larger and deeper. As a result, the effort needed to train a network also increases dramatically. In many cases, it is more practical to use a neural network intellectual property (IP) that an IP vendor has already trained. As we do not know about the training process, there can be security threats in the neural IP: the IP vendor (attacker) may embed hidden malicious functionality, i.e. neural Trojans, into the neural IP. We show that this is an effective attack and provide three mitigation techniques: input anomaly detection, re-training, and input preprocessing. All the techniques are proven effective. The input anomaly detection approach is able to detect 99.8% of Trojan triggers although with 12.2% false positive. The re-training approach is able to prevent 94.1% of Trojan triggers from triggering the Trojan although it requires that the neural IP be reconfigurable. In the input preprocessing approach, 90.2% of Trojan triggers are rendered ineffective and no assumption about the neural IP is needed.

cs.CR↗

arXiv2017

Power Optimizations in MTJ-based Neural Networks through Stochastic Computing

Artificial Neural Networks (ANNs) have found widespread applications in tasks such as pattern recognition and image classification. However, hardware implementations of ANNs using conventional binary arithmetic units are computationally expensive, energy-intensive and have large area overheads. Stochastic Computing (SC) is an emerging paradigm which replaces these conventional units with simple logic circuits and is particularly suitable for fault-tolerant applications. Spintronic devices, such as Magnetic Tunnel Junctions (MTJs), are capable of replacing CMOS in memory and logic circuits. In this work, we propose an energy-efficient use of MTJs, which exhibit probabilistic switching behavior, as Stochastic Number Generators (SNGs), which forms the basis of our NN implementation in the SC domain. Further, error resilient target applications of NNs allow us to introduce Approximate Computing, a framework wherein accuracy of computations is traded-off for substantial reductions in power consumption. We propose approximating the synaptic weights in our MTJ-based NN implementation, in ways brought about by properties of our MTJ-SNG, to achieve energy-efficiency. We design an algorithm that can perform such approximations within a given error tolerance in a single-layer NN in an optimal way owing to the convexity of the problem formulation. We then use this algorithm and develop a heuristic approach for approximating multi-layer NNs. To give a perspective of the effectiveness of our approach, a 43% reduction in power consumption was obtained with less than 1% accuracy loss on a standard classification problem, with 26% being brought about by the proposed algorithm.

cs.NE↗