Search

This work introduces Gemma, a family of lightweight, state-of-the art open models built from the research and technology used to create Gemini models. Gemma models demonstrate strong performance across academic benchmarks for language understanding, reasoning, and safety. We release two sizes of models (2 billion and 7 billion parameters), and provide both pretrained and fine-tuned checkpoints. Gemma outperforms similarly sized open models on 11 out of 18 text-based tasks, and we present comprehensive evaluations of safety and responsibility aspects of the models, alongside a detailed description of model development. We believe the responsible release of LLMs is critical for improving the safety of frontier models, and for enabling the next wave of LLM innovations.

Time crystals are many-body states that spontaneously break translation symmetry in time the way that ordinary crystals do in space. While experimental observations have confirmed the existence of discrete or continuous time crystals, these realizations have relied on the utilization of periodic forces or effective modulation through cavity feedback. The original proposal for time crystals is that they would represent self-sustained motions without any external periodicity, but realizing such purely self-generated behavior has not yet been achieved. Here, we provide theoretical and experimental evidence that many-body interactions can give rise to an inherent time crystalline phase. Following a calculation that shows an ensemble of pumped four-level atoms can spontaneously break continuous time translation symmetry, we observe periodic motions in an erbium-doped solid. The inherent time crystal produced by our experiment is self-protected by many-body interactions and has a measured coherence time beyond that of individual erbium ions.

Large Language Models (LLMs) have demonstrated impressive capabilities for text rewriting. Nonetheless, the large sizes of these models make them impractical for on-device inference, which would otherwise allow for enhanced privacy and economical inference. Creating a smaller yet potent language model for text rewriting presents a formidable challenge because it requires balancing the need for a small size with the need to retain the emergent capabilities of the LLM, that requires costly data collection. To address the above challenge, we introduce a new instruction tuning approach for building a mobile-centric text rewriting model. Our strategies enable the generation of high quality training data without any human labeling. In addition, we propose a heuristic reinforcement learning framework which substantially enhances performance without requiring preference data. To further bridge the performance gap with the larger server-side model, we propose an effective approach that combines the mobile rewrite agent with the server model using a cascade. To tailor the text rewriting tasks to mobile scenarios, we introduce MessageRewriteEval, a benchmark that focuses on text rewriting for messages through natural language instructions. Through empirical experiments, we demonstrate that our on-device model surpasses the current state-of-the-art LLMs in text rewriting while maintaining a significantly reduced model size. Notably, we show that our proposed cascading approach improves model performance.

The rapid development and application of foundation models have revolutionized the field of artificial intelligence. Large diffusion models have gained significant attention for their ability to generate photorealistic images and support various tasks. On-device deployment of these models provides benefits such as lower server costs, offline functionality, and improved user privacy. However, common large diffusion models have over 1 billion parameters and pose challenges due to restricted computational and memory resources on devices. We present a series of implementation optimizations for large diffusion models that achieve the fastest reported inference latency to-date (under 12 seconds for Stable Diffusion 1.4 without int8 quantization on Samsung S23 Ultra for a 512x512 image with 20 iterations) on GPU-equipped mobile devices. These enhancements broaden the applicability of generative AI and improve the overall user experience across a wide range of devices., Comment: 4 pages (not including references), 2 figures, 2 tables. Accepted to Efficient Deep Learning for Computer Vision workshop 2023

Integrated optical quantum memories are a scalable solution to synchronize a large number of quantum nodes. Without compact quantum memories, some astonishing quantum applications such as distributed quantum computing and quantum sensor networks would not be possible. Rather than find a specific material that meets all the requirements of an on-chip quantum memory as other protocols usually do, we propose to assign the memory requirements on coherent storage and controllability to rare earth ions and a lithium niobate crystal, respectively. Specifically, optical quantum states are stored in an erbium-doped lithium niobate micro-cavity by utilizing the electro-optic effect of lithium niobate. The cavity frequency can be shifted by an external electric field, thus modifying the resonance condition between the cavity and the collective atomic excitation. This effect is further used to suppress or enhance the emission of photon echoes. Our calculated results show that high efficiency and low noise performance is achievable.

Integrated optics provides a platform for the experimental implementation of highly complex and compact circuits for practical applications as well as for advances in the fundamental science of quantum optics. The lithium niobate (LN) waveguide is an important candidate for the construction of integrated optical circuits. Based on the bound state in the continuum (BIC) in a LN waveguide, we propose an efficient way to produce polarization-entangled photon pairs. The implementation of this method is simple and does not require the polarization process needed for periodically poled LN. The generation rate of the entangled photon pairs increases linearly with the length of the waveguide. For visible light, the generation efficiency can be improved by more than five orders of magnitude with waveguides having the length of only a few millimeters, compared with the corresponding case without BICs. The phenomena can appear in a very wide spectrum range from the visible to THz regions. This study is of great significance for the development of active integrated quantum chips in various wavelength ranges.

Bandgap control is of central importance for semiconductor technologies. The traditional means of control is to dope the lattice chemically, electrically or optically with charge carriers. Here, we demonstrate for the first time a widely tunable bandgap (renormalisation up to 650 meV at room-temperature) in two-dimensional (2D) semiconductors by coherently doping the lattice with plasmonic hot electrons. In particular, we integrate tungsten-disulfide (WS$_2$) monolayers into a self-assembled plasmonic crystal, which enables coherent coupling between semiconductor excitons and plasmon resonances. Accompanying this process, the plasmon-induced hot electrons can repeatedly fill the WS$_2$ conduction band, leading to population inversion and a significant reconstruction in band structures and exciton relaxations. Our findings provide an innovative and effective measure to engineer optical responses of 2D semiconductors, allowing a great flexiblity in design and optimisation of photonic and optoelectronic devices.

We demonstrate an optical unstable phase for a laser driven erbium ion ensemble. The transmitted light through an erbium-doped yttrium orthosillicate crystal becomes dynamically unstable when illuminated by a strong continuous-wave laser. Transient net gain was recorded if the light passes the sample twice. The experimental results are understood in the framework of a many-body system interacting with a classical field, where the ion-ion interaction becomes significant as a result of the high erbium concentration. A Bloch-equation model that includes the excitation-induced frequency-shift is introduced to discuss the conditions of the instabilities., Comment: 5 pages, 4 figures

Strong coupling of two-dimensional semiconductor excitons with plasmonic resonators enables control of light-matter interaction at the subwavelength scale. Here we develop strong coupling in plasmonic nano-gap resonators that allow modification of exciton number contributing to the coupling. Using this system, we not only demonstrate a large vacuum Rabi splitting up to 163 meV and splitting features in photoluminescence spectra, but also reveal that the exciton number can be reduced down to single-digit level (N<10), which is an order lower than that of traditional systems, close to single-exciton based strong coupling. In addition, we prove that the strong coupling process is not affected by the large exciton coherence size that was previously believed to be detrimental to the formation of plasmon-exciton interaction. Our work provides a deeper understanding of storng coupling in two-dimensional semiconductors, paving the way for room temperature quantum optics applications.

We show that crystal-field calculations for $C_1$ point-group symmetry are possible, and that such calculations can be performed with sufficient accuracy to have substantial utility for rare-earth based quantum information applications. In particular, we perform crystal-field fitting for a C$_1$-symmetry site in $^{167}$Er$^{3+}$:Y$_2$SiO$_5$. The calculation simultaneously includes site-selective spectroscopic data up to 20,000 cm$^{-1}$, rotational Zeeman data, and ground- and excited-state hyperfine structure determined from high-resolution Raman-heterodyne spectroscopy on the 1.5 $\mu$m telecom transition. We achieve an agreement of better than 50 MHz for assigned hyperfine transitions. The success of this analysis opens the possibility of systematically evaluating the coherence properties, as well as transition energies and intensities, of any rare-earth ion doped into Y$_2$SiO$_5$ ., Comment: 6 pages, plus 5 pages in supplementary information, 4 figures total

Spins in solids are an ideal candidate to act as a memory and interface with superconducting qubits due to their long coherence times. We spectroscopically investigate erbium-167-doped yttrium orthosilicate as a possible microwave-addressed memory employing its microwave frequency transitions that occur without applying an external magnetic field. We obtain coherence times of 380 $\mu$s in a ground state spin transition and 1.48 ms in an excited state spin transition. This is 28 times longer compared to previous zero field measurements, as well as 200 times longer than a previous microwave memory demonstration in the same material. These long coherence times show that erbium-167-doped yttrium orthosilicate has potential as a microwave-addressed quantum memory., Comment: 9 pages, 7 figures. The paper has been expanded compared to the previous version on arXiv, and the title has changed

The efficiency of the frequency conversion process at the heart of Raman heterodyne spectroscopy was improved by nearly four orders of magnitude by resonant enhancement of both the pump and signal optical fields. Our results using an erbium doped Y$_2$SiO$_5$ crystal at temperatures near 4K suggest that such an approach is promising for the quantum conversion of microwave to optical photons., Comment: 5 pages, 4 figures

The hyperfine structure of the ground state of erbium doped yttrium orthosilicate is analyzed with the use of electron paramagnetic resonance experiments in a tunable microwave resonator. This work was prompted by the disagreement between the measurements made in zero magnetic field [Phys. Rev. B, 94, 075117, (2016)] and a previously published spin Hamiltonian, which is derived from standard EPR measurements at 9.5 GHz [Phys. Rev. B, 74, 214409, (2006)]. The ability to vary magnetic field strength, resonator frequency, and the orientation of our sample enabled us to monitor how the frequencies of hyperfine transitions change as a function of a vector magnetic field. Compared to [Phys. Rev. B, 74, 214409, (2006)], we arrived at a different set of spin Hamiltonian parameters, which are also broadly consistent with their data. We discuss the reliability of our new spin Hamiltonian parameters to make predictions outside the magnetic field and frequency regimes of our data. We also discuss why it proved to be difficult to determine spin Hamiltonian parameters for this material, and present data collection strategies that improve the model reliability., Comment: Supporting Information of this paper is available

Here we demonstrate the gas sensing ability of cavity-coupled metallic nanoparticle systems, comprising gold nanoparticles separated from a gold mirror with a polymer spacer. An increase in relative humidity (RH) causes the spacer to expand, which induces a significant reduction of nanoparticle scattering intensity, as the scattering is highly dependent on the cavity-nanoparticle coupling that closely relates to the nanoparticle-mirror distance. This lithography-free structure enables a remarkable averaging sensitivity at 0.12 dB/% RH and 0.25 dB/% RH over RH range (45-75%), possessing an estimated resolution better than 0.5% RH with full reversibility and almost zero-hysteresis, exhibiting notable gas sensing potentials.

Plasmonic cavities can be used to control the atom-photon coupling process at the nanoscale, since they provide ultrahigh density of optical states in an exceptionally small mode volume. Here we demonstrate strong coupling between molecular excitons and plasmonic resonances (so-called plexcitonic coupling) in a film-coupled nanocube cavity, which can induce profound and significant spectral and spatial modifications to the plasmonic gap modes. Within the spectral span of a single gap mode in the nanotube-film cavity with a 3-nm wide gap, the introduction of narrow-band J-aggregate dye molecules not only enables an anti-crossing behavior in the spectral response, but also splits the single spatial mode into two distinct modes that are easily identified by their far-field scattering profiles. Simulation results confirm the experimental findings and the sensitivity of the plexcitonic coupling is explored using digital control of the gap spacing. Our work opens up a new perspective to study the strong coupling process, greatly extending the functionality of nanophotonic systems, with the potential to be applied in cavity quantum electrodynamic systems.

We experimentally demonstrate the coupling at zero magnetic field of an isotopically pure erbium doped yttrium orthosilicate crystal ($^{167}$Er:YSO) to a three-dimensional superconducting cavity with a $Q$ factor of $10^5$. A tunable loop-gap resonator is used, and its resonance frequency is tuned to observe the hyperfine transitions of the erbium sample. The observed spectrum differs from what is predicted by the published spin Hamiltonian parameters. The narrow cavity linewidth also enables the observation of asymmetric line shapes for these hyperfine transitions. Such a broadly tunable superconducting cavity (from 1.6 GHz to 4.0 GHz in the current design) is a promising device for building hybrid quantum systems., Comment: 5 figures

Head movement during scanning impedes activation detection in fMRI studies. Head motion in fMRI acquired using slice-based Echo Planar Imaging (EPI) can be estimated and compensated by aligning the images onto a reference volume through image registration. However, registering EPI images volume to volume fails to consider head motion between slices, which may lead to severely biased head motion estimates. Slice-to-volume registration can be used to estimate motion parameters for each slice by more accurately representing the image acquisition sequence. However, accurate slice to volume mapping is dependent on the information content of the slices: middle slices are information rich, while edge slides are information poor and more prone to distortion. In this work, we propose a Gaussian particle filter based head motion tracking algorithm to reduce the image misregistration errors. The algorithm uses a dynamic state space model of head motion with an observation equation that models continuous slice acquisition of the scanner. Under this model the particle filter provides more accurate motion estimates and voxel position estimates. We demonstrate significant performance improvement of the proposed approach as compared to registration-only methods of head motion estimation and brain activation detection., Comment: This paper has been submitted to Transaction on Medical Imaging

We treat the problem of estimation of orientation parameters whose values are invariant to transformations from a spherical symmetry group. Previous work has shown that any such group-invariant distribution must satisfy a restricted finite mixture representation, which allows the orientation parameter to be estimated using an Expectation Maximization (EM) maximum likelihood (ML) estimation algorithm. In this paper, we introduce two parametric models for this spherical symmetry group estimation problem: 1) the hyperbolic Von Mises Fisher (VMF) mixture distribution and 2) the Watson mixture distribution. We also introduce a new EM-ML algorithm for clustering samples that come from mixtures of group-invariant distributions with different parameters. We apply the models to the problem of mean crystal orientation estimation under the spherically symmetric group associated with the crystal form, e.g., cubic or octahedral or hexahedral. Simulations and experiments establish the advantages of the extended EM-VMF and EM-Watson estimators for data acquired by Electron Backscatter Diffraction (EBSD) microscopy of a polycrystalline Nickel alloy sample., Comment: arXiv admin note: text overlap with arXiv:1411.2540

We propose a framework for indexing of grain and sub-grain structures in electron backscatter diffraction (EBSD) images of polycrystalline materials. The framework is based on a previously introduced physics-based forward model by Callahan and De Graef (2013) relating measured patterns to grain orientations (Euler angle). The forward model is tuned to the microscope and the sample symmetry group. We discretize the domain of the forward model onto a dense grid of Euler angles and for each measured pattern we identify the most similar patterns in the dictionary. These patterns are used to identify boundaries, detect anomalies, and index crystal orientations. The statistical distribution of these closest matches is used in an unsupervised binary decision tree (DT) classifier to identify grain boundaries and anomalous regions. The DT classifies a pattern as an anomaly if it has an abnormally low similarity to any pattern in the dictionary. It classifies a pixel as being near a grain boundary if the highly ranked patterns in the dictionary differ significantly over the pixels 3x3 neighborhood. Indexing is accomplished by computing the mean orientation of the closest dictionary matches to each pattern. The mean orientation is estimated using a maximum likelihood approach that models the orientation distribution as a mixture of Von Mises-Fisher distributions over the quaternionic 3-sphere. The proposed dictionary matching approach permits segmentation, anomaly detection, and indexing to be performed in a unified manner with the additional benefit of uncertainty quantification. We demonstrate the proposed dictionary-based approach on a Ni-base IN100 alloy., Comment: This paper is in press in the Journal of Microscopy and Microanalysis, Cambridge University Press, Feb. 2015