Search results

We study the problem of differentially private (DP) fine-tuning of large pre-trained models — a recent privacy-preserving approach suitable for solving downstream tasks with sensitive data. Existing work has demonstrated that high accuracy is possible under strong privacy constraint, yet requires significant computational overhead or modifications to the network architecture. We propose differentially private

Programmable arrays of hundreds of Rydberg atoms have recently enabled the exploration of remarkable phenomena in many-body quantum physics. In addition, the development of high-fidelity quantum gates are making them promising architectures for the implementation of quantum circuits. We present here Pulser, an open-source Python library for programming neutral-atom devices at the pulse level. The low-level

Achieving an accurate description of fermionic systems typically requires considerably many more orbitals than fermions. Previous resource analyses of quantum chemistry simulation often failed to exploit this low fermionic number information in the implementation of Trotter-based approaches and overestimated the quantum-computer runtime as a result. They also depended on numerical procedures that are computationally

Simulation of the Hubbard model is a leading candidate for the first useful applications of a fault-tolerant quantum computer. A recent study of quantum algorithms for early simulations of the Hubbard model [Kivlichan et al 2019 Quantum 4 296] found that the lowest resource costs were achieved by split-operator Trotterization combined with the fast-fermionic Fourier transform (FFFT) on an L × L lattice

A highly anticipated use of quantum computers is the simulation of complex quantum systems including molecules and other many-body systems. One promising method involves directly applying a linear combination of unitaries (LCU) to approximate a Taylor series by truncating after some order. Here we present an adaptation of that method, optimized for Hamiltonians with terms of widely varying magnitude, as

Hyperfine atomic states are among the most promising candidates for qubit encoding in quantum information processing. In atomic systems, hyperfine transitions are typically driven through a two-photon Raman process by a laser field which is amplitude modulated at the hyperfine qubit frequency. Here we introduce a method for generating amplitude modulation by phase modulating a laser and reflecting it from

Heatsinks may cause radiated emission and radio frequency interference problems when they are mounted on printed circuit boards. In this paper, the radiation mechanism of heatsinks is systematically investigated using characteristic mode theory. The dipole moment is a commonly used equivalent source model for integrated circuits that drive radiated emission from heatsinks. On the basis of a simplified modal

Code generation models have achieved impressive performance. However, they tend to be brittle as slight edits to a prompt could lead to very different generations; these robustness properties, critical for user experience when deployed in real-life applications, are not well understood. Most existing works on robustness in text or code tasks have focused on classification, while robustness in generation

Measurements of a physical quantity by measuring devices are usually noisy enough that we need to correct, or at least mitigate, the effects of noise. For this purpose, it’s important to distinguish between systematic and random noise since they are of a different nature and independent from each other (when defined properly), so should be dealt with differently. For example, random noise can be significantly

Protected qubits such as the 0-π qubit, and bosonic qubits including cat qubits and Gottesman-Kitaev-Preskill (GKP) qubits offer advantages for fault tolerance. Some of these protected qubits (e.g., 0-π qubit and Kerr-cat qubit) are stabilized by Hamiltonians which have (near-)degenerate ground state manifolds with large energy gaps to the excited state manifolds. Without dissipative stabilization mechanisms

Kernel Stein discrepancies (KSDs) are maximum mean discrepancies (MMDs) that leverage the score information of distributions, and have grown central to a wide range of applications. In most settings, these MMDs are required to (i) separate a target P from other probability measures or even (ii) control weak convergence to P. In this article we derive new sufficient and necessary conditions that substantially

There is an increasing trend in using neural methods for dialogue model evaluation. Lack of a framework to investigate these metrics can cause dialogue models to reflect their biases and cause unforeseen problems during interactions. In this work, we propose an adversarial test-suite which generates problematic variations of various dialogue aspects, e.g. logical entailment, using automatic heuristics.

In many speech-enabled human-machine interactions, user speech can overlap with the device playback audio. In these instances, the performance of tasks such as keyword-spotting (KWS) and device-directed speech detection (DDD) can de- grade significantly. To address this problem, we propose an implicit acoustic echo cancellation (iAEC) framework where a neural network is trained to exploit the additional

Synthesis of many-body quantum systems in the laboratory can provide further insight into the emergent behavior of quantum materials. While the majority of engineerable many-body systems, or quantum simulators, consist of particles on a lattice with local interactions, quantum systems featuring long-range interactions are particularly difficult to model and interesting to study due to the rapid spatio-temporal

We show that the proof of the generalised quantum Stein's lemma [Brandão & Plenio, Commun. Math. Phys. 295, 791 (2010)] is not correct due to a gap in the argument leading to Lemma III.9. Hence, the main achievability result of Brandão & Plenio is not known to hold. This puts into question a number of established results in the literature, in particular the reversibility of quantum entanglement [Brandão

Controlling long-lived mechanical oscillators in the quantum regime holds promises for quantum information processing. Here, we present an electromechanical system capable of operating in the GHz-frequency band in a silicon-on-insulator platform. Relying on a novel driving scheme based on an electrostatic field and high-impedance microwave cavities based on TiN superinductors, we are able to demonstrate

We demonstrate a superconducting artificial atom with strong unidirectional coupling to a microwave photonic waveguide. Our artificial atom is realized by coupling a transmon qubit to the waveguide at two spatially separated points with time-modulated interactions. Direction-sensitive interference arising from the parametric couplings in our scheme results in a non-reciprocal response, where we measure

Dynamic graph forecasting has found a wide range of applications including social media, recommendation systems, and computational finance. However, existing dynamic graph models typically focus on discrete-time dynamic graphs, treating dynamic graphs as temporally discrete graph snapshots. We argue that such discrete treatment is inadequate for capturing the underlying dynamics which are intrinsically

Due to intense interest in the potential applications of quantum computing, it is critical to understand the basis for potential exponential quantum advantage in quantum chemistry. Here we gather the evidence for this case in the most common task in quantum chemistry, namely, ground-state energy estimation, for generic chemical problems where heuristic quantum state preparation might be assumed to be efficient

The goal of meta-learning is to learn to adapt to a new task with only a few labeled examples. Inspired by the recent progress in large language models, we propose in-context tuning (ICT), which recasts task adaptation and prediction as a simple sequence prediction problem: to form the input sequence, we concatenate the task instruction, labeled in-context examples, and the target input to predict; to metatrain