A5048472359- Quantum Computing Algorithms and Architecture
- Topic Modeling
- Natural Language Processing Techniques
- Speech and dialogue systems
- Quantum-Dot Cellular Automata
- Quantum and electron transport phenomena
- Quantum Information and Cryptography
- Machine Learning in Materials Science
- Coding theory and cryptography
- Text Readability and Simplification
- Neural Networks and Reservoir Computing
- Molecular Communication and Nanonetworks
- Advanced Graph Neural Networks
- Computability, Logic, AI Algorithms
- Multimodal Machine Learning Applications
- Advanced biosensing and bioanalysis techniques
- Digital Image Processing Techniques
- Complexity and Algorithms in Graphs
- Interconnection Networks and Systems
- Software Engineering Research
- Brain Tumor Detection and Classification
- Advanced Data Storage Technologies
- Blind Source Separation Techniques
- Evolutionary Algorithms and Applications
- Neural Networks and Applications
IBM Research - India
2020-2024
IBM (United States)
2017-2020
The University of Sydney
1980
Quantum Approximate Optimization Algorithm (QAOA) is studied primarily to find approximate solutions combinatorial optimization problems. For a graph with $n$ vertices and $m$ edges, depth $p$ QAOA for the Max-cut problem requires $2\cdot m \cdot p$ CNOT gates. one of primary sources error in modern quantum computers. In this paper, we propose two hardware independent methods reduce number gates circuit. First, present method based on Edge Coloring input that minimizes cycles (termed as...
In recent years, there has been tremendous progress in the development of quantum computing hardware, algorithms and services leading to expectation that near future computers will be capable performing simulations for natural science applications, operations research, machine learning at scales mostly inaccessible classical computers. Whereas impact already started recognized fields such as cryptanalysis, simulations, optimization among others, very little is known about full potential...
Neural Conversational QA tasks such as ShARC require systems to answer questions based on the contents of a given passage. On studying recent state-of-the-art models task, we found indications that model(s) learn spurious clues/patterns in data-set. Further, heuristic-based program, built exploit these patterns, had comparative performance neural models. In this paper share our findings about four types patterns corpus and how them. Motivated by above findings, create modified data-set has...
While a Quantum Approximate Optimization Algorithm (QAOA) is intended to provide quantum advantage in finding approximate solutions combinatorial optimization problems, noise the system hurdle exploiting its full potential. Several error mitigation techniques have been studied lessen effect of on this algorithm. Recently, Majumdar et al. proposed Depth First Search (DFS) based method reduce $n-1$ CNOT gates ansatz design QAOA for Max-Cut graph G = (V, E), |V| n. However, tends increase depth...
There are several dialog frameworks which allow manual specification of intents and rule based flow. The framework provides good control to designers at the expense being more time consuming laborious. job a designer can be reduced if we could identify pairs user corresponding responses automatically from prior conversations between users agents. In this paper propose an approach find these frequent utterances (which serve as examples for intents) agent responses. We novel SimCluster...
Quantum Approximate Optimization Algorithm (QAOA) is a prospective candidate for providing quantum advantage in finding approximate solutions to optimization problems using near-term devices. In this paper, the goal reduce number of CNOT gates and depth ansatz circuits QAOA. First, we present generalized QAOA formulation any Hamiltonian that involves upto two-body interactions, as graph. The circuit realization depth-p requires $2mp$ where graph has n vertices m edges. Presently, gate one...
Advances in classical machine learning and single-cell technologies have paved the way to understand interactions between disease cells tumor microenvironments accelerate therapeutic discovery. However, challenges these methods NP-hard problems spatial Biology create an opportunity for quantum computing algorithms. We a hybrid quantum-classical graph neural network (GNN) that combines GNN with Variational Quantum Classifier (VQC) classifying binary sub-tasks breast cancer subtyping. explore...
Errors in heavy hexagonal code and other topological codes like surface were usually decoded using the Minimum Weight Perfect Matching (MWPM) based decoders. Recent advances have shown that can be efficiently by deploying machine learning (ML) techniques, for example, neural networks. In this work, we first propose an ML decoder show decode efficiently, terms of values threshold pseudo-threshold, various noise models. We proposed decoding method achieves $\sim 5$ times higher than MWPM....
An important task in quantum generative machine learning is to model the probability distribution of measurements many-body systems. Classical models, such as adversarial networks (GANs) and variational autoencoders (VAEs), can distributions product states with high fidelity, but fail or require an exponential number parameters entangled states. In this paper, we introduce a quantum-enhanced VAE (QeVAE), quantum-classical hybrid that uses correlations improve fidelity over classical VAEs,...
Error syndromes for heavy hexagonal code and other topological codes such as surface have typically been decoded by using Minimum Weight Perfect Matching– (MWPM) based methods. Recent advances shown that can be efficiently deploying machine learning (ML) techniques, in particular with neural networks. In this work, we first propose an ML-based decoder establish its efficiency terms of the values threshold pseudo-threshold various noise models. We show proposed decoding method achieves ~ 5 ×...
Recently several deep learning based models have been proposed for end-to-end of dialogs. While these can be trained from data without the need any additional annotations, it is hard to interpret them. On other hand, there exist traditional state dialog systems, where states are discrete and hence easy interpret. However handcrafted annotated in data. To achieve best both worlds, we propose Latent State Tracking Network (LSTN) using which learn an interpretable model unsupervised manner. The...
Neural Conversational QA tasks like ShARC require systems to answer questions based on the contents of a given passage. On studying recent state-of-the-art models ShARCQA task, we found indications that learn spurious clues/patterns in dataset. Furthermore, show heuristic-based program designed exploit these patterns can have performance comparable neural models. In this paper share our findings about four types corpus and describe how them. Motivated by aforementioned findings, create...