Search

NL2SQL (Natural Language to Structured Query Language) transformation has seen wide adoption in Business Intelligence (BI) applications in recent years. However, existing NL2SQL benchmarks are not suitable for production BI scenarios, as they are not designed for common business intelligence questions. To address this gap, we have developed a new benchmark focused on typical NL questions in industrial BI scenarios. We discuss the challenges of constructing a BI-focused benchmark and the shortcomings of existing benchmarks. Additionally, we introduce question categories in our benchmark that reflect common BI inquiries. Lastly, we propose two novel semantic similarity evaluation metrics for assessing NL2SQL capabilities in BI applications and services., Comment: This paper has been accepted by ICSOC (International Conference on Service-Oriented Computing) 2024

Measuring Efficiency in neural network system development is an open research problem. This paper presents an experimental framework to measure the training efficiency of a neural architecture. To demonstrate our approach, we analyze the training efficiency of Convolutional Neural Networks and Bayesian equivalents on the MNIST and CIFAR-10 tasks. Our results show that training efficiency decays as training progresses and varies across different stopping criteria for a given neural model and learning task. We also find a non-linear relationship between training stopping criteria, training Efficiency, model size, and training Efficiency. Furthermore, we illustrate the potential confounding effects of overtraining on measuring the training efficiency of a neural architecture. Regarding relative training efficiency across different architectures, our results indicate that CNNs are more efficient than BCNNs on both datasets. More generally, as a learning task becomes more complex, the relative difference in training efficiency between different architectures becomes more pronounced.

Density matrices are the most general descriptions of quantum states, covering both pure and mixed states. Positive semidefiniteness is a physical requirement of density matrices, imposing nonnegative probabilities of measuring physical values. Separately, nonlocality is a property shared by some bipartite quantum systems, indicating a correlation of the component parts that cannot be described by local classical variables. In this work, we show that breaking the positive-semidefinite requirement and allowing states with a negative minimal eigenvalue arbitrarily close to zero, allows for the construction of states that are nonlocal under one component labelling but local when the labelling is interchanged. This is an observer-dependent nonlocality, showing the connection between nonlocal objectivism and negative quasi-probabilities., Comment: An error in the calculation of the eigenvalues for \beta^{T}\beta for the right-identity case (Sec. 3, eq. 8 line 1). In fact the eigenvalues are the same for both left- and right- identities, leaving the main result null and void

We introduce and describe a new heuristic method for finding an upper bound on the degree of contextuality and the corresponding unsatisfied part of a quantum contextual configuration with three-element contexts (i.e., lines) located in a multi-qubit symplectic polar space of order two. While the previously used method based on a SAT solver was limited to three qubits, this new method is much faster and more versatile, enabling us to also handle four- to six-qubit cases. The four-qubit unsatisfied configurations we found are quite remarkable. That of an elliptic quadric features 315 lines and has in its core three copies of the split Cayley hexagon of order two having a Heawood-graph-underpinned geometry in common. That of a hyperbolic quadric also has 315 lines but, as a point-line incidence structure, is isomorphic to the dual $\mathcal{DW}(5,2)$ of $\mathcal{W}(5,2)$. Finally, an unsatisfied configuration with 1575 lines associated with all the lines/contexts of the four-qubit space contains a distinguished $\mathcal{DW}(5,2)$ centered on a point-plane incidence graph of PG$(3,2)$. The corresponding configurations found in the five-qubit space exhibit a considerably higher degree of complexity, except for a hyperbolic quadric, whose 6975 unsatisfied contexts are compactified around the point-hyperplane incidence graph of PG$(4,2)$. The most remarkable unsatisfied patterns discovered in the six-qubit space are a couple of disjoint split Cayley hexagons (for the full space) and a subgeometry underpinned by the complete bipartite graph $K_{7,7}$ (for a hyperbolic quadric)., Comment: 35 pages, 14 figures

This study presents a novel methodology incorporating safety constraints into a robotic simulation during the training of deep reinforcement learning (DRL). The framework integrates specific parts of the safety requirements, such as velocity constraints, as specified by ISO 10218, directly within the DRL model that becomes a part of the robot's learning algorithm. The study then evaluated the efficiency of these safety constraints by subjecting the DRL model to various scenarios, including grasping tasks with and without obstacle avoidance. The validation process involved comprehensive simulation-based testing of the DRL model's responses to potential hazards and its compliance. Also, the performance of the system is carried out by the functional safety standards IEC 61508 to determine the safety integrity level. The study indicated a significant improvement in the safety performance of the robotic system. The proposed DRL model anticipates and mitigates hazards while maintaining operational efficiency. This study was validated in a testbed with a collaborative robotic arm with safety sensors and assessed with metrics such as the average number of safety violations, obstacle avoidance, and the number of successful grasps. The proposed approach outperforms the conventional method by a 16.5% average success rate on the tested scenarios in the simulations and 2.5% in the testbed without safety violations. The project repository is available at https://github.com/ammar-n-abbas/sim2real-ur-gym-gazebo., Comment: This paper has been accepted for publication in the proceedings of the IEEE/IFAC International Conference on Control, Decision, and Information Technologies (CoDIT), 2024

We demonstrate a vacuum-gap ultrastable optical reference cavity that does not require a vacuum enclosure. Our simple method of optical contact bonding in a vacuum environment allows for cavity operation in air while maintaining vacuum between the cavity mirrors. Vacuum is maintained long term, with no observed degradation in cavity stability for over 1 year after bonding. For a 1550 nm laser stabilized to a 9.7 mL in-vacuum bonded cavity, the measured Allan deviation is $2.4\times 10^{-14}$ at 1 s and its phase noise is thermal-noise-limited from 0.1 Hz to 10 kHz, reaching about -105 dBc/Hz at 10 kHz offset frequency. This represents the highest stability of any oscillator operated without a vacuum enclosure. Furthermore, we demonstrate a 0.5 mL in-vacuum bonded cavity created using microfabricated mirrors and cavity dicing, with phase noise reaching -95 dBc/Hz at 10 kHz offset frequency. By relieving the need for high-vacuum enclosures, we greatly enhance the portability and utility of low noise, compact cavity-stabilized lasers, with applications ranging from environmental sensing to mobile optical clocks to ultralow noise microwave generation., Comment: 10 pages, 6 figures

Stroke is one of the leading causes of death and disability worldwide but it is believed to be highly preventable. The majority of stroke prevention focuses on targeting high-risk individuals but its is important to understand how the targeting of high-risk individuals might impact the overall societal burden of stroke. We propose using an agent-based model that follows agents through their pre-stroke and stroke journey to assess the impacts of different interventions at the population level. We present a case study looking at the impacts of agents being informed of their stroke risk at certain ages and those agents taking measure to reduce their risk. The results of our study show that if agents are aware of their risk and act accordingly we see a significant reduction in strokes and population DALYs. The case study highlights the importance of individuals understanding their own stroke risk for stroke prevention and the usefulness of agent-based models in assessing the impact of stroke interventions., Comment: 18th Social Simulation Conference

We study the mass distribution of galaxy clusters in Milgromian dynamics, or modified Newtonian dynamics (MOND). We focus on five galaxy clusters from the X-COP sample, for which high-quality data are available on both the baryonic mass distribution (gas and stars) and internal dynamics (from the hydrostatic equilibrium of hot gas and the Sunyaev-Zeldovich effect). We confirm that galaxy clusters require additional `missing matter' in MOND, although the required amount is drastically reduced with respect to the non-baryonic dark matter in the context of Newtonian dynamics. We studied the spatial distribution of the missing matter by fitting the acceleration profiles of the clusters with a Bayesian method, finding that a physical density profile with an inner core and an outer $r^{-4}$ decline (giving a finite total mass) provide good fits within $\sim$1 Mpc. At larger radii, the fit results are less satisfactory but the combination of the MOND external field effect and hydrostatic bias (quantified as 10$\%$-40$\%$) can play a key role. The missing mass must be more centrally concentrated than the intracluster medium (ICM). For relaxed clusters (A1795, A2029, A2142), the ratio of missing-to-visible mass is around $1-5$ at $R\simeq200-300$ kpc and decreases to $0.4-1.1$ at $R\simeq2-3$ Mpc, showing that the total amount of missing mass is smaller than or comparable to the ICM mass. For clusters with known merger signatures (A644 and A2319), this global ratio increases up to $\sim$5 but may indicate out-of-equilibrium dynamics rather than actual missing mass. We discuss various possibilities regarding the nature of the extra mass, in particular `missing baryons' in the form of pressure-confined cold gas clouds with masses of $<10^5$ M$_\odot$ and sizes of $< 50$ pc., Comment: Accepted for publication in A&A

Next-generation communications, radar and navigation systems will extend and exploit the higher bandwidth of the millimeter-wave domain for increased communication data rates as well as radar with higher sensitivity and increased spatial resolution. However, realizing these advantages will require the generation of millimeter-wave signals with low phase noise in simple and compact form-factors. The rapidly developing field of photonic integration addresses this challenge and provides a path toward simplified and portable, low-noise mm-wave generation for these applications. We leverage these advances by heterodyning two silicon photonic chip lasers, phase-locked to the same miniature Fabry-Perot (F-P) cavity to demonstrate a simple framework for generating low-noise millimeter-waves with phase noise below the thermal limit of the F-P cavity. Specifically, we generate 94.5 GHz and 118.1 GHz millimeter-wave signals with phase noise of -117 dBc/Hz at 10 kHz offset, decreasing to -120 dBc/Hz at 40 kHz offset, a record low value for such photonic devices. We achieve this with existing technologies that can be integrated into a platform less than $\approx$ 10 mL in volume. Our work illustrates the significant potential and advantages of low size, weight, and power (SWaP) photonic-sourced mm-waves for communications and sensing., Comment: 5 pages, 4 figures

Mass customization and shorter manufacturing cycles are becoming more important among small and medium-sized companies. However, classical industrial robots struggle to cope with product variation and dynamic environments. In this paper, we present CoBT, a collaborative programming by demonstration framework for generating reactive and modular behavior trees. CoBT relies on a single demonstration and a combination of data-driven machine learning methods with logic-based declarative learning to learn a task, thus eliminating the need for programming expertise or long development times. The proposed framework is experimentally validated on 7 manipulation tasks and we show that CoBT achieves approx. 93% success rate overall with an average of 7.5s programming time. We conduct a pilot study with non-expert users to provide feedback regarding the usability of CoBT., Comment: Accepted for presentation at IEEE ICRA 2024

Modern communication, navigation, and radar systems rely on low noise and frequency-agile microwave sources. In this application space, photonic systems provide an attractive alternative to conventional microwave synthesis by leveraging high spectral purity lasers and optical frequency combs to generate microwaves with exceedingly low phase noise. However, these photonic techniques suffer from a lack of frequency tunability, and also have substantial size, weight, and power requirements that largely limit their use to laboratory settings. In this work, we address these shortcomings with a hybrid opto-electronic approach that combines simplified optical frequency division with direct digital synthesis to produce tunable low-phase-noise microwaves across the entire X-band. This results in exceptional phase noise at 10 GHz of -156 dBc/Hz at 10 kHz offset and fractional frequency instability of 1x10^-13 at 0.1 s. Spot tuning away from 10 GHz by 500 MHz, 1 GHz, and 2 GHz, yields phase noise at 10 kHz offset of -150 dBc/Hz, -146 dBc/Hz, and -140 dBc/Hz, respectively. The synthesizer architecture is fully compatible with integrated photonic implementations that will enable a versatile microwave source in a chip-scale package. Together, these advances illustrate an impactful and practical synthesis technique that shares the combined benefits of low timing noise provided by photonics and the frequency agility of established digital synthesis.

This study focuses on the novel application of a normalizing flow as a method of domain adaptation. Normalizing flows offer a way to transform data points between two different distributions. The present study investigates a method of transforming latent representations of physics data to a normal distribution and then to a physics distribution again. The final distribution models a simulated distribution. Following the transformation process, the data can be classified by a neural network trained on labeled simulation data. The present study succeeds in training two normalizing flows that can transform between data (or simulation) and a Gaussian distribution., Comment: To be published in JINST as part of proceedings for AI4EIC2023. 6 pages, 3 figures

The present study presents a novel application for normalizing flows for domain adaptation. The study investigates the ability of flow based neural networks to improve signal extraction of $\Lambda$ Hyperons at CLAS12. Normalizing Flows can help model complex probability density functions that describe physics processes, enabling uses such as event generation. $\Lambda$ signal extraction has been improved through the use of classifier networks, but differences in simulation and data domains limit classifier performance; this study utilizes the flows for domain adaptation between Monte Carlo simulation and data. We were successful in training a flow network to transform between the latent physics space and a normal distribution. We also found that applying the flows lessened the dependence of the figure of merit on the cut on the classifier output, meaning that there was a broader range where the cut results in a similar figure of merit., Comment: Proceedings for the 25th International Spin Physics Symposium (SPIN 2023)

Quantum games embody non-intuitive consequences of quantum phenomena, such as entanglement and contextuality. The Mermin-Peres game is a simple example, demonstrating how two players can utilise shared quantum information to win a no - communication game with certainty, where classical players cannot. In this paper we look at the geometric structure behind such classical strategies, and borrow ideas from the geometry of symplectic polar spaces to maximise this quantum advantage. We introduce a new game called the Eloily game with a quantum-classical success gap of $0.2\overline{6}$, larger than that of the Mermin-Peres and doily games. We simulate this game in the IBM Quantum Experience and obtain a success rate of $1$, beating the classical bound of $0.7\overline{3}$ demonstrating the efficiency of the quantum strategy., Comment: 22 pages, 8 figures, 2 tables, 1 appendix

Transformer-based Neural Language Models achieve state-of-the-art performance on various natural language processing tasks. However, an open question is the extent to which these models rely on word-order/syntactic or word co-occurrence/topic-based information when processing natural language. This work contributes to this debate by addressing the question of whether these models primarily use topic as a signal, by exploring the relationship between Transformer-based models' (BERT and RoBERTa's) performance on a range of probing tasks in English, from simple lexical tasks such as sentence length prediction to complex semantic tasks such as idiom token identification, and the sensitivity of these tasks to the topic information. To this end, we propose a novel probing method which we call topic-aware probing. Our initial results indicate that Transformer-based models encode both topic and non-topic information in their intermediate layers, but also that the facility of these models to distinguish idiomatic usage is primarily based on their ability to identify and encode topic. Furthermore, our analysis of these models' performance on other standard probing tasks suggests that tasks that are relatively insensitive to the topic information are also tasks that are relatively difficult for these models.

In complex industrial and chemical process control rooms, effective decision-making is crucial for safety and efficiency. The experiments in this paper evaluate the impact and applications of an AI-based decision support system integrated into an improved human-machine interface, using dynamic influence diagrams, a hidden Markov model, and deep reinforcement learning. The enhanced support system aims to reduce operator workload, improve situational awareness, and provide different intervention strategies to the operator adapted to the current state of both the system and human performance. Such a system can be particularly useful in cases of information overload when many alarms and inputs are presented all within the same time window, or for junior operators during training. A comprehensive cross-data analysis was conducted, involving 47 participants and a diverse range of data sources such as smartwatch metrics, eye-tracking data, process logs, and responses from questionnaires. The results indicate interesting insights regarding the effectiveness of the approach in aiding decision-making, decreasing perceived workload, and increasing situational awareness for the scenarios considered. Additionally, the results provide valuable insights to compare differences between styles of information gathering when using the system by individual participants. These findings are particularly relevant when predicting the overall performance of the individual participant and their capacity to successfully handle a plant upset and the alarms connected to it using process and human-machine interaction logs in real-time. These predictions enable the development of more effective intervention strategies.

In this paper we introduce TWIG (Topologically-Weighted Intelligence Generation), a novel, embedding-free paradigm for simulating the output of KGEs that uses a tiny fraction of the parameters. TWIG learns weights from inputs that consist of topological features of the graph data, with no coding for latent representations of entities or edges. Our experiments on the UMLS dataset show that a single TWIG neural network can predict the results of state-of-the-art ComplEx-N3 KGE model nearly exactly on across all hyperparameter configurations. To do this it uses a total of 2590 learnable parameters, but accurately predicts the results of 1215 different hyperparameter combinations with a combined cost of 29,322,000 parameters. Based on these results, we make two claims: 1) that KGEs do not learn latent semantics, but only latent representations of structural patterns; 2) that hyperparameter choice in KGEs is a deterministic function of the KGE model and graph structure. We further hypothesise that, as TWIG can simulate KGEs without embeddings, that node and edge embeddings are not needed to learn to accurately predict new facts in KGs. Finally, we formulate all of our findings under the umbrella of the ``Structural Generalisation Hypothesis", which suggests that ``twiggy" embedding-free / data-structure-based learning methods can allow a single neural network to simulate KGE performance, and perhaps solve the Link Prediction task, across many KGs from diverse domains and with different semantics., Comment: This article was accepted for publication at IEEE ICSC 2024

As it is well known, split Cayley hexagons of order two live in the three-qubit symplectic polar space in two non-isomorphic embeddings, called classical and skew. Although neither of the two embeddings yields observable-based contextual configurations of their own, {\it classically}-embedded copies are found to fully rule contextuality properties of the most prominent three-qubit contextual configurations in the following sense: each set of unsatisfiable contexts of such a contextual configuration is isomorphic to the set of lines that certain classically-embedded hexagon shares with this particular configuration. In particular, for a doily this shared set comprises three pairwise disjoint lines belonging to a grid of the doily, for an elliptic quadric the corresponding set features nine mutually disjoint lines forming a (Desarguesian) spread on the quadric, for a hyperbolic quadric the set entails 21 lines that are in bijection with the edges of the Heawood graph and, finally, for the configuration that consists of all the 315 contexts of the space its 63 unsatisfiable ones cover an entire hexagon. A particular illustration of this encoding is provided by the {\it line-complement} of a skew-embedded hexagon; its 24 unsatisfiable contexts correspond exactly to those 24 lines in which a particular classical copy of the hexagon differs from the considered skew-embedded one. In connection with the last-mentioned case we also conducted some experimental tests on a Noisy Intermediate Scale Quantum (NISQ) computer to validate our theoretical findings., Comment: 25 pages, 14 figures

Micro-channel plate (MCP) detectors, when used at pulsed-neutron-source instruments, offer the possibility of high spatial resolution and high contrast imaging with pixel-level spectroscopic information. Here we demonstrate the possibility of multimodal analysis including total neutron cross-section spectra measurements, quantitative material differentiation imaging and texture-sensitive in-line phase imaging, from a single exposure using an MCP detector. This multimodal approach operates in full-field imaging mode, with the neutron transmission spectra acquired at each individual detector pixel. Due to the polychromatic nature of the beam and spectroscopic resolving capability of the detector, no energy scanning is required. Good agreement with the library reference data is demonstrated for neutron cross-section spectra measurements. Two different images corresponding to two selected energy bandwidths are used for elemental differentiation imaging. Moreover, the presence of changes in texture, i.e., preferred grain orientation, in the sample is identified from our phase-retrieval imaging results., Comment: Revised manuscript resubmitted to Physical Review A

The difficulty of identifying the physical model of complex systems has led to exploring methods that do not rely on such complex modeling of the systems. Deep reinforcement learning has been the pioneer for solving this problem without the need for relying on the physical model of complex systems by just interacting with it. However, it uses a black-box learning approach that makes it difficult to be applied within real-world and safety-critical systems without providing explanations of the actions derived by the model. Furthermore, an open research question in deep reinforcement learning is how to focus the policy learning of critical decisions within a sparse domain. This paper proposes a novel approach for the use of deep reinforcement learning in safety-critical systems. It combines the advantages of probabilistic modeling and reinforcement learning with the added benefits of interpretability and works in collaboration and synchronization with conventional decision-making strategies. The BC-SRLA is activated in specific situations which are identified autonomously through the fused information of probabilistic model and reinforcement learning, such as abnormal conditions or when the system is near-to-failure. Further, it is initialized with a baseline policy using policy cloning to allow minimum interactions with the environment to address the challenges associated with using RL in safety-critical industries. The effectiveness of the BC-SRLA is demonstrated through a case study in maintenance applied to turbofan engines, where it shows superior performance to the prior art and other baselines., Comment: arXiv admin note: text overlap with arXiv:2206.13433

This paper discusses the methods that we used for our submissions to the WMT 2023 Terminology Shared Task for German-to-English (DE-EN), English-to-Czech (EN-CS), and Chinese-to-English (ZH-EN) language pairs. The task aims to advance machine translation (MT) by challenging participants to develop systems that accurately translate technical terms, ultimately enhancing communication and understanding in specialised domains. To this end, we conduct experiments that utilise large language models (LLMs) for two purposes: generating synthetic bilingual terminology-based data, and post-editing translations generated by an MT model through incorporating pre-approved terms. Our system employs a four-step process: (i) using an LLM to generate bilingual synthetic data based on the provided terminology, (ii) fine-tuning a generic encoder-decoder MT model, with a mix of the terminology-based synthetic data generated in the first step and a randomly sampled portion of the original generic training data, (iii) generating translations with the fine-tuned MT model, and (iv) finally, leveraging an LLM for terminology-constrained automatic post-editing of the translations that do not include the required terms. The results demonstrate the effectiveness of our proposed approach in improving the integration of pre-approved terms into translations. The number of terms incorporated into the translations of the blind dataset increases from an average of 36.67% with the generic model to an average of 72.88% by the end of the process. In other words, successful utilisation of terms nearly doubles across the three language pairs., Comment: WMT 2023

Traditional controllers have limitations as they rely on prior knowledge about the physics of the problem, require modeling of dynamics, and struggle to adapt to abnormal situations. Deep reinforcement learning has the potential to address these problems by learning optimal control policies through exploration in an environment. For safety-critical environments, it is impractical to explore randomly, and replacing conventional controllers with black-box models is also undesirable. Also, it is expensive in continuous state and action spaces, unless the search space is constrained. To address these challenges we propose a specialized deep residual policy safe reinforcement learning with a cycle of learning approach adapted for complex and continuous state-action spaces. Residual policy learning allows learning a hybrid control architecture where the reinforcement learning agent acts in synchronous collaboration with the conventional controller. The cycle of learning initiates the policy through the expert trajectory and guides the exploration around it. Further, the specialization through the input-output hidden Markov model helps to optimize policy that lies within the region of interest (such as abnormality), where the reinforcement learning agent is required and is activated. The proposed solution is validated on the Tennessee Eastman process control.

The nanoscale contrast in scattering-type scanning near-field optical microscopy (s-SNOM) is determined by the optical properties of the sample immediately under the apex of the tip of the atomic force microscope (AFM). There are several models that describe the optical scattering of an incident field by the tip near a surface, and these models have been successful in relating the measured scattering signal to the dielectric function of the sample under the tip. Here, we address a situation that is normally not considered in the existing interaction models, namely the near-field signal arising from thin, highly conductive films in the terahertz (THz) frequency range. According to established theoretical models, highly conductive thin films should show insignificant contrast in the THz range for small variations in conductivity, therefore hindering the use of s-SNOM for nanoscale characterisation. We experimentally demonstrate unexpected but clear and quantifiable layer contrast in the THz s-SNOM signal from few-layer exfoliated graphene as well as subtle nanoscale contrast variations within graphene layers. We use finite-element simulations to confirm that the observed contrast is described by the classical electromagnetics of the scattering mechanism, suggesting that the dipole models must be reformulated to correctly describe the interaction with conductive samples.

It is known that Mermin-Peres like proofs of quantum contextuality can furnish non-local games with a guaranteed quantum strategy, when classically no such guarantee can exist. This phenomenon, also called quantum pseudo-telepathy, has been studied in the case of the so-called Mermin Magic square game. In this paper we review in detail two different ways of implementing on a quantum computer such a game and propose a new Doily game based on the geometry of 2-qubit Pauli group. We show that the quantumness of these games are almost revealed when we play them on the IBM Quantum Experience, however the inherent noise in the available quantum machines prevents a full demonstration of the non-classical aspects., Comment: 16 pages, 7 figures, 4 tables. For associated codes, see https://quantcert.github.io/quantum_game

Increasing the average power of broadband, few-cycle terahertz (THz) sources is currently a topic of intense investigation, fueled by recent immense progress in high average power femtosecond laser driving sources at 1030 nm. However, many crucial applications would benefit not only from an increase in average power, but also from ultra-broad bandwidth, while maintaining high dynamic range at these frequencies. This calls for the challenging combination of high repetition rates and high average power simultaneously. Here, we discuss the recent progress in the promising approach enabled by organic crystals for THz-generation. Specifically, this review article discusses advances with the most commonly used organic crystals BNA, DAST, DSTMS, OH1 and HMQ-TMS. We place special emphasis on nonlinear and thermal properties and discuss future directions for this field.

Numerous modern technologies are reliant on the low-phase noise and exquisite timing stability of microwave signals. Substantial progress has been made in the field of microwave photonics, whereby low noise microwave signals are generated by the down-conversion of ultra-stable optical references using a frequency comb. Such systems, however, are constructed with bulk or fiber optics and are difficult to further reduce in size and power consumption. Our work addresses this challenge by leveraging advances in integrated photonics to demonstrate low-noise microwave generation via two-point optical frequency division. Narrow linewidth self-injection locked integrated lasers are stabilized to a miniature Fabry-P\'{e}rot cavity, and the frequency gap between the lasers is divided with an efficient dark-soliton frequency comb. The stabilized output of the microcomb is photodetected to produce a microwave signal at 20 GHz with phase noise of -96 dBc/Hz at 100 Hz offset frequency that decreases to -135 dBc/Hz at 10 kHz offset--values which are unprecedented for an integrated photonic system. All photonic components can be heterogeneously integrated on a single chip, providing a significant advance for the application of photonics to high-precision navigation, communication and timing systems.

We present an in-vacuum bonded, 9.7 mL-volume Fabry-P\'erot ultrastable optical reference cavity that operates without a vacuum enclosure. A laser stabilized to the cavity demonstrates low, thermal noise-limited phase noise and 5x10^{-14} Allan deviation at 1 second., Comment: Submitted to IEEE Photonics Conference 2023

Environmental monitoring is crucial to our understanding of climate change, biodiversity loss and pollution. The availability of large-scale spatio-temporal data from sources such as sensors and satellites allows us to develop sophisticated models for forecasting and understanding key drivers. However, the data collected from sensors often contain missing values due to faulty equipment or maintenance issues. The missing values rarely occur simultaneously leading to data that are multivariate misaligned sparse time series. We propose two models that are capable of performing multivariate spatio-temporal forecasting while handling missing data naturally without the need for imputation. The first model is a transformer-based model, which we name SERT (Spatio-temporal Encoder Representations from Transformers). The second is a simpler model named SST-ANN (Sparse Spatio-Temporal Artificial Neural Network) which is capable of providing interpretable results. We conduct extensive experiments on two different datasets for multivariate spatio-temporal forecasting and show that our models have competitive or superior performance to those at the state-of-the-art., Comment: 11 pages, 7 figures