arXiv subjects

Bharat Kaul

Publications and source records attributed to Bharat Kaul.

At least 19 recordsLinked to original sources

arXiv2026

Scalable Pretraining of Large Mixture of Experts Language Models on Aurora Super Computer

Pretraining Large Language Models (LLMs) from scratch requires massive amount of compute. Aurora super computer is an ExaScale machine with 127,488 Intel PVC (Ponte Vechio) GPU tiles. In this work, we showcase LLM pretraining on Aurora at the scale of 1000s of GPU tiles. Towards this effort, we developed Optimus, an inhouse training library with support for standard large model training techniques. Using Optimus, we first pretrained Mula-1B, a 1 Billion dense model and Mula-7B-A1B, a 7 Billion Mixture of Experts (MoE) model from scratch on 3072 GPU tiles for the full 4 trillion tokens of the OLMoE-mix-0924 dataset. We then demonstrated model scaling by pretraining three large MoE models Mula-20B-A2B, Mula-100B-A7B, and Mula-220B-A10B till 100 Billion tokens on the same dataset. On our largest model Mula-220B-A10B, we pushed the compute scaling from 384 to 12288 GPU tiles and observed scaling efficiency of around 90% at 12288 GPU tiles. We significantly improved the runtime performance of MoE models using custom GPU kernels for expert computation, and a novel EP-Aware sharded optimizer resulting in training speedups up to 1.71x. As part of the Optimus library, we also developed a robust set of reliability and fault tolerant features to improve training stability and continuity at scale.

cs.LG↗

arXiv2025

Multi-modal encoder-decoder neural network for forecasting solar wind speed at L1

The solar wind, accelerated within the solar corona, sculpts the heliosphere and continuously interacts with planetary atmospheres. On Earth, high-speed solar-wind streams may lead to severe disruption of satellite operations and power grids. Accurate and reliable forecasting of the ambient solar-wind speed is therefore highly desirable. This work presents an encoder-decoder neural-network framework for simultaneously forecasting the daily averaged solar-wind speed for the subsequent four days. The encoder-decoder framework is trained with the two different modes of solar observations. The history of solar-wind observations from prior solar-rotations and EUV coronal observations up to four days prior to the current time form the input to two different encoders. The decoder is designed to output the daily averaged solar-wind speed from four days prior to the current time to four days into the future. Our model outputs the solar-wind speed with Root-Mean-Square Errors (RMSEs) of 55 km/s, 58 km/s, 58 km/s, and 58 km/s and Pearson correlations of 0.78, 0.66, 0.64 and 0.63 for one to four days in advance respectively. While the model is trained and validated on observations between 2010 - 2018, we demonstrate its robustness via application on unseen test data between 2019 - 2023, yielding RMSEs of 53 km/s and Pearson correlations 0.55 for a four-day advance prediction. Our encoder-decoder model thus produces much improved RMSE values compared to the previous works and paves the way for developing comprehensive multimodal deep learning models for operational solar wind forecasting.

astro-ph.SR↗

arXiv2024

Generative Active Learning for the Search of Small-molecule Protein Binders

Despite substantial progress in machine learning for scientific discovery in recent years, truly de novo design of small molecules which exhibit a property of interest remains a significant challenge. We introduce LambdaZero, a generative active learning approach to search for synthesizable molecules. Powered by deep reinforcement learning, LambdaZero learns to search over the vast space of molecules to discover candidates with a desired property. We apply LambdaZero with molecular docking to design novel small molecules that inhibit the enzyme soluble Epoxide Hydrolase 2 (sEH), while enforcing constraints on synthesizability and drug-likeliness. LambdaZero provides an exponential speedup in terms of the number of calls to the expensive molecular docking oracle, and LambdaZero de novo designed molecules reach docking scores that would otherwise require the virtual screening of a hundred billion molecules. Importantly, LambdaZero discovers novel scaffolds of synthesizable, drug-like inhibitors for sEH. In in vitro experimental validation, a series of ligands from a generated quinazoline-based scaffold were synthesized, and the lead inhibitor N-(4,6-di(pyrrolidin-1-yl)quinazolin-2-yl)-N-methylbenzamide (UM0152893) displayed sub-micromolar enzyme inhibition of sEH.

q-bio.BM↗

arXiv2023

AUTOSPARSE: Towards Automated Sparse Training of Deep Neural Networks

Sparse training is emerging as a promising avenue for reducing the computational cost of training neural networks. Several recent studies have proposed pruning methods using learnable thresholds to efficiently explore the non-uniform distribution of sparsity inherent within the models. In this paper, we propose Gradient Annealing (GA), where gradients of masked weights are scaled down in a non-linear manner. GA provides an elegant trade-off between sparsity and accuracy without the need for additional sparsity-inducing regularization. We integrated GA with the latest learnable pruning methods to create an automated sparse training algorithm called AutoSparse, which achieves better accuracy and/or training/inference FLOPS reduction than existing learnable pruning methods for sparse ResNet50 and MobileNetV1 on ImageNet-1K: AutoSparse achieves (2x, 7x) reduction in (training,inference) FLOPS for ResNet50 on ImageNet at 80% sparsity. Finally, AutoSparse outperforms sparse-to-sparse SotA method MEST (uniform sparsity) for 80% sparse ResNet50 with similar accuracy, where MEST uses 12% more training FLOPS and 50% more inference FLOPS.

cs.LG↗

arXiv2022

Measuring frequency and period separations in red-giant stars using machine learning

Asteroseismology is used to infer the interior physics of stars. The \textit{Kepler} and TESS space missions have provided a vast data set of red-giant light curves, which may be used for asteroseismic analysis. These data sets are expected to significantly grow with future missions such as \textit{PLATO}, and efficient methods are therefore required to analyze these data rapidly. Here, we describe a machine learning algorithm that identifies red giants from the raw oscillation spectra and captures \textit{p} and \textit{mixed} mode parameters from the red-giant power spectra. We report algorithmic inferences for large frequency separation ($Δν$), frequency at maximum amplitude ($ν_{max}$), and period separation ($ΔΠ$) for an ensemble of stars. In addition, we have discovered $\sim$25 new probable red giants among 151,000 \textit{Kepler} long-cadence stellar-oscillation spectra analyzed by the method, among which four are binary candidates which appear to possess red-giant counterparts. To validate the results of this method, we selected $\sim$ 3,000 \textit{Kepler} stars, at various evolutionary stages ranging from subgiants to red clumps, and compare inferences of $Δν$, $ΔΠ$, and $ν_{max}$ with estimates obtained using other techniques. The power of the machine-learning algorithm lies in its speed: it is able to accurately extract seismic parameters from 1,000 spectra in $\sim$5 seconds on a modern computer (single core of the Intel Xeon Platinum 8280 CPU).

astro-ph.SR↗

arXiv2021

SEERL: Sample Efficient Ensemble Reinforcement Learning

Ensemble learning is a very prevalent method employed in machine learning. The relative success of ensemble methods is attributed to their ability to tackle a wide range of instances and complex problems that require different low-level approaches. However, ensemble methods are relatively less popular in reinforcement learning owing to the high sample complexity and computational expense involved in obtaining a diverse ensemble. We present a novel training and model selection framework for model-free reinforcement algorithms that use ensembles of policies obtained from a single training run. These policies are diverse in nature and are learned through directed perturbation of the model parameters at regular intervals. We show that learning and selecting an adequately diverse set of policies is required for a good ensemble while extreme diversity can prove detrimental to overall performance. Selection of an adequately diverse set of policies is done through our novel policy selection framework. We evaluate our approach on challenging discrete and continuous control tasks and also discuss various ensembling strategies. Our framework is substantially sample efficient, computationally inexpensive and is seen to outperform state-of-the-art (SOTA) scores in Atari 2600 and Mujoco.

cs.LG↗

arXiv2021

Efficient and Generic 1D Dilated Convolution Layer for Deep Learning

Convolutional neural networks (CNNs) have found many applications in tasks involving two-dimensional (2D) data, such as image classification and image processing. Therefore, 2D convolution layers have been heavily optimized on CPUs and GPUs. However, in many applications - for example genomics and speech recognition, the data can be one-dimensional (1D). Such applications can benefit from optimized 1D convolution layers. In this work, we introduce our efficient implementation of a generic 1D convolution layer covering a wide range of parameters. It is optimized for x86 CPU architectures, in particular, for architectures containing Intel AVX-512 and AVX-512 BFloat16 instructions. We use the LIBXSMM library's batch-reduce General Matrix Multiplication (BRGEMM) kernel for FP32 and BFloat16 precision. We demonstrate that our implementation can achieve up to 80% efficiency on Intel Xeon Cascade Lake and Cooper Lake CPUs. Additionally, we show the generalization capability of our BRGEMM based approach by achieving high efficiency across a range of parameters. We consistently achieve higher efficiency than the 1D convolution layer with Intel oneDNN library backend for varying input tensor widths, filter widths, number of channels, filters, and dilation parameters. Finally, we demonstrate the performance of our optimized 1D convolution layer by utilizing it in the end-to-end neural network training with real genomics datasets and achieve up to 6.86x speedup over the oneDNN library-based implementation on Cascade Lake CPUs. We also demonstrate the scaling with 16 sockets of Cascade/Cooper Lake CPUs and achieve significant speedup over eight V100 GPUs using a similar power envelop. In the end-to-end training, we get a speedup of 1.41x on Cascade Lake with FP32, 1.57x on Cooper Lake with FP32, and 2.27x on Cooper Lake with BFloat16 over eight V100 GPUs with FP32.

cs.LG↗

arXiv2021

AI Powered Compiler Techniques for DL Code Optimization

Creating high performance implementations of deep learning primitives on CPUs is a challenging task. Multiple considerations including multi-level cache hierarchy, and wide SIMD units of CPU platforms influence the choice of program transformations to apply for performance optimization. In this paper, we present machine learning powered compiler techniques to optimize loop nests. We take a two-pronged approach to code optimization: We first apply high level optimizations to optimize the code to take optimal advantage of the cache memories. Then, we perform low level, target-specific optimizations to effectively vectorize the code to run well on the SIMD units of the machine. For high level optimizations, we use polyhedral compilation techniques and deep learning approaches. For low level optimization, we use a target specific code generator that generates code using vector intrinsics and Reinforcement Learning (RL) techniques to find the optimal parameters for the code generator. We perform experimental evaluation of the developed techniques on various matrix multiplications that occur in popular deep learning workloads. The experimental results show that the compiler techniques presented in the paper achieve 7.6X and 8.2X speed-ups over a baseline for sequential and parallel runs respectively.

cs.PL↗

arXiv2021

GNNerator: A Hardware/Software Framework for Accelerating Graph Neural Networks

Graph Neural Networks (GNNs) use a fully-connected layer to extract features from the nodes of a graph and aggregate these features using message passing between nodes, combining two distinct computational patterns: dense, regular computations and sparse, irregular computations. To address this challenge, we propose GNNerator, an accelerator with heterogeneous compute engines optimized for these two patterns. Further, GNNerator implements feature-blocking, a novel GNN dataflow that beneficially trades off irregular memory accesses during aggregation for regular memory accesses during feature extraction. We show GNNerator achieves speedups of 5.7-37x over an NVIDIA RTX 2080-Ti, and 2.3x-3.8x over HyGCN, a state-of-the-art GNN accelerator.

cs.AR↗

arXiv2020

PolyDL: Polyhedral Optimizations for Creation of High Performance DL primitives

Deep Neural Networks (DNNs) have revolutionized many aspects of our lives. The use of DNNs is becoming ubiquitous including in softwares for image recognition, speech recognition, speech synthesis, language translation, to name a few. he training of DNN architectures however is computationally expensive. Once the model is created, its use in the intended application - the inference task, is computationally heavy too and the inference needs to be fast for real time use. For obtaining high performance today, the code of Deep Learning (DL) primitives optimized for specific architectures by expert programmers exposed via libraries is the norm. However, given the constant emergence of new DNN architectures, creating hand optimized code is expensive, slow and is not scalable. To address this performance-productivity challenge, in this paper we present compiler algorithms to automatically generate high performance implementations of DL primitives that closely match the performance of hand optimized libraries. We develop novel data reuse analysis algorithms using the polyhedral model to derive efficient execution schedules automatically. In addition, because most DL primitives use some variant of matrix multiplication at their core, we develop a flexible framework where it is possible to plug in library implementations of the same in lieu of a subset of the loops. We show that such a hybrid compiler plus a minimal library-use approach results in state-of-the-art performance. We develop compiler algorithms to also perform operator fusions that reduce data movement through the memory hierarchy of the computer system.

cs.DC↗

arXiv2020

MADRaS : Multi Agent Driving Simulator

In this work, we present MADRaS, an open-source multi-agent driving simulator for use in the design and evaluation of motion planning algorithms for autonomous driving. MADRaS provides a platform for constructing a wide variety of highway and track driving scenarios where multiple driving agents can train for motion planning tasks using reinforcement learning and other machine learning algorithms. MADRaS is built on TORCS, an open-source car-racing simulator. TORCS offers a variety of cars with different dynamic properties and driving tracks with different geometries and surface properties. MADRaS inherits these functionalities from TORCS and introduces support for multi-agent training, inter-vehicular communication, noisy observations, stochastic actions, and custom traffic cars whose behaviours can be programmed to simulate challenging traffic conditions encountered in the real world. MADRaS can be used to create driving tasks whose complexities can be tuned along eight axes in well-defined steps. This makes it particularly suited for curriculum and continual learning. MADRaS is lightweight and it provides a convenient OpenAI Gym interface for independent control of each car. Apart from the primitive steering-acceleration-brake control mode of TORCS, MADRaS offers a hierarchical track-position -- speed control that can potentially be used to achieve better generalization. MADRaS uses multiprocessing to run each agent as a parallel process for efficiency and integrates well with popular reinforcement learning libraries like RLLib.

cs.RO↗

arXiv2020

K-TanH: Efficient TanH For Deep Learning

We propose K-TanH, a novel, highly accurate, hardware efficient approximation of popular activation function TanH for Deep Learning. K-TanH consists of parameterized low-precision integer operations, such as, shift and add/subtract (no floating point operation needed) where parameters are stored in very small look-up tables that can fit in CPU registers. K-TanH can work on various numerical formats, such as, Float32 and BFloat16. High quality approximations to other activation functions, e.g., Sigmoid, Swish and GELU, can be derived from K-TanH. Our AVX512 implementation of K-TanH demonstrates $>5\times$ speed up over Intel SVML, and it is consistently superior in efficiency over other approximations that use floating point arithmetic. Finally, we achieve state-of-the-art Bleu score and convergence results for training language translation model GNMT on WMT16 data sets with approximate TanH obtained via K-TanH on BFloat16 inputs.

cs.LG↗

arXiv2020

PolyScientist: Automatic Loop Transformations Combined with Microkernels for Optimization of Deep Learning Primitives

At the heart of deep learning training and inferencing are computationally intensive primitives such as convolutions which form the building blocks of deep neural networks. Researchers have taken two distinct approaches to creating high performance implementations of deep learning kernels, namely, 1) library development exemplified by Intel MKL-DNN for CPUs, 2) automatic compilation represented by the TensorFlow XLA compiler. The two approaches have their drawbacks: even though a custom built library can deliver very good performance, the cost and time of development of the library can be high. Automatic compilation of kernels is attractive but in practice, till date, automatically generated implementations lag expert coded kernels in performance by orders of magnitude. In this paper, we develop a hybrid solution to the development of deep learning kernels that achieves the best of both worlds: the expert coded microkernels are utilized for the innermost loops of kernels and we use the advanced polyhedral technology to automatically tune the outer loops for performance. We design a novel polyhedral model based data reuse algorithm to optimize the outer loops of the kernel. Through experimental evaluation on an important class of deep learning primitives namely convolutions, we demonstrate that the approach we develop attains the same levels of performance as Intel MKL-DNN, a hand coded deep learning library.

cs.PL↗

arXiv2019

High Performance Scalable FPGA Accelerator for Deep Neural Networks

Low-precision is the first order knob for achieving higher Artificial Intelligence Operations (AI-TOPS). However the algorithmic space for sub-8-bit precision compute is diverse, with disruptive changes happening frequently, making FPGAs a natural choice for Deep Neural Network inference, In this work we present an FPGA-based accelerator for CNN inference acceleration. We use {\it INT-8-2} compute (with {\it 8 bit} activation and {2 bit} weights) which is recently showing promise in the literature, and which no known ASIC, CPU or GPU natively supports today. Using a novel Adaptive Logic Module (ALM) based design, as a departure from traditional DSP based designs, we are able to achieve high performance measurement of 5 AI-TOPS for {\it Arria10} and project a performance of 76 AI-TOPS at 0.7 TOPS/W for {\it Stratix10}. This exceeds known CPU, GPU performance and comes close to best known ASIC (TPU) numbers, while retaining the versatility of the FPGA platform for other applications.

cs.DC↗

arXiv2019

A Study of BFLOAT16 for Deep Learning Training

This paper presents the first comprehensive empirical study demonstrating the efficacy of the Brain Floating Point (BFLOAT16) half-precision format for Deep Learning training across image classification, speech recognition, language modeling, generative networks and industrial recommendation systems. BFLOAT16 is attractive for Deep Learning training for two reasons: the range of values it can represent is the same as that of IEEE 754 floating-point format (FP32) and conversion to/from FP32 is simple. Maintaining the same range as FP32 is important to ensure that no hyper-parameter tuning is required for convergence; e.g., IEEE 754 compliant half-precision floating point (FP16) requires hyper-parameter tuning. In this paper, we discuss the flow of tensors and various key operations in mixed precision training, and delve into details of operations, such as the rounding modes for converting FP32 tensors to BFLOAT16. We have implemented a method to emulate BFLOAT16 operations in Tensorflow, Caffe2, IntelCaffe, and Neon for our experiments. Our results show that deep learning training using BFLOAT16 tensors achieves the same state-of-the-art (SOTA) results across domains as FP32 tensors in the same number of iterations and with no changes to hyper-parameters.

cs.LG↗

arXiv2019

Automatic Model Parallelism for Deep Neural Networks with Compiler and Hardware Support

The deep neural networks (DNNs) have been enormously successful in tasks that were hitherto in the human-only realm such as image recognition, and language translation. Owing to their success the DNNs are being explored for use in ever more sophisticated tasks. One of the ways that the DNNs are made to scale for the complex undertakings is by increasing their size -- deeper and wider networks can model well the additional complexity. Such large models are trained using model parallelism on multiple compute devices such as multi-GPUs and multi-node systems. In this paper, we develop a compiler-driven approach to achieve model parallelism. We model the computation and communication costs of a dataflow graph that embodies the neural network training process and then, partition the graph using heuristics in such a manner that the communication between compute devices is minimal and we have a good load balance. The hardware scheduling assistants are proposed to assist the compiler in fine tuning the distribution of work at runtime.

cs.DC↗

arXiv2019

Mixed Precision Training With 8-bit Floating Point

Reduced precision computation for deep neural networks is one of the key areas addressing the widening compute gap driven by an exponential growth in model size. In recent years, deep learning training has largely migrated to 16-bit precision, with significant gains in performance and energy efficiency. However, attempts to train DNNs at 8-bit precision have met with significant challenges because of the higher precision and dynamic range requirements of back-propagation. In this paper, we propose a method to train deep neural networks using 8-bit floating point representation for weights, activations, errors, and gradients. In addition to reducing compute precision, we also reduced the precision requirements for the master copy of weights from 32-bit to 16-bit. We demonstrate state-of-the-art accuracy across multiple data sets (imagenet-1K, WMT16) and a broader set of workloads (Resnet-18/34/50, GNMT, Transformer) than previously reported. We propose an enhanced loss scaling method to augment the reduced subnormal range of 8-bit floating point for improved error propagation. We also examine the impact of quantization noise on generalization and propose a stochastic rounding technique to address gradient noise. As a result of applying all these techniques, we report slightly higher validation accuracy compared to full precision baseline.

cs.LG↗

arXiv2018

Out-of-Distribution Detection Using an Ensemble of Self Supervised Leave-out Classifiers

As deep learning methods form a critical part in commercially important applications such as autonomous driving and medical diagnostics, it is important to reliably detect out-of-distribution (OOD) inputs while employing these algorithms. In this work, we propose an OOD detection algorithm which comprises of an ensemble of classifiers. We train each classifier in a self-supervised manner by leaving out a random subset of training data as OOD data and the rest as in-distribution (ID) data. We propose a novel margin-based loss over the softmax output which seeks to maintain at least a margin $m$ between the average entropy of the OOD and in-distribution samples. In conjunction with the standard cross-entropy loss, we minimize the novel loss to train an ensemble of classifiers. We also propose a novel method to combine the outputs of the ensemble of classifiers to obtain OOD detection score and class prediction. Overall, our method convincingly outperforms Hendrycks et al.[7] and the current state-of-the-art ODIN[13] on several OOD detection benchmarks.

cs.LG↗