Search

Efficient decoding of quantum codes is crucial for achieving high-performance quantum error correction. In this paper, we introduce the concept of approximate degenerate decoding and integrate it with ordered statistics decoding (OSD). Previously, we proposed a reliability metric that leverages both hard and soft decisions from the output of belief propagation (BP), which is particularly useful for identifying highly reliable subsets of variables. Using the approach of reliable subset reduction, we reduce the effective problem size. Additionally, we identify a degeneracy condition that allows high-order OSD to be simplified to order-0 OSD. By integrating these techniques, we present an ADOSD algorithm that significantly improves OSD efficiency in the code capacity noise model. We demonstrate the effectiveness of our BP+ADOSD approach through extensive simulations on a varity of quantum codes, including generalized hypergraph-product codes, topological codes, lift-connected surface codes, and bivariate bicycle codes. The results indicate that the BP+ADOSD decoder outperforms existing methods, achieving higher error thresholds and enhanced performance at low error rates. Additionally, we validate the efficiency of our approach in terms of computational time, demonstrating that ADOSD requires, on average, the same amount of time as two to three BP iterations on surface codes at a depolarizing error rate of around $1\%$. All the proposed algorithms are compared using single-threaded CPU implementations., Comment: 16 pages, 10 figures, 6 tables

Photorealistic 3D scene reconstruction plays an important role in autonomous driving, enabling the generation of novel data from existing datasets to simulate safety-critical scenarios and expand training data without additional acquisition costs. Gaussian Splatting (GS) facilitates real-time, photorealistic rendering with an explicit 3D Gaussian representation of the scene, providing faster processing and more intuitive scene editing than the implicit Neural Radiance Fields (NeRFs). While extensive GS research has yielded promising advancements in autonomous driving applications, they overlook two critical aspects: First, existing methods mainly focus on low-speed and feature-rich urban scenes and ignore the fact that highway scenarios play a significant role in autonomous driving. Second, while LiDARs are commonplace in autonomous driving platforms, existing methods learn primarily from images and use LiDAR only for initial estimates or without precise sensor modeling, thus missing out on leveraging the rich depth information LiDAR offers and limiting the ability to synthesize LiDAR data. In this paper, we propose a novel GS method for dynamic scene synthesis and editing with improved scene reconstruction through LiDAR supervision and support for LiDAR rendering. Unlike prior works that are tested mostly on urban datasets, to the best of our knowledge, we are the first to focus on the more challenging and highly relevant highway scenes for autonomous driving, with sparse sensor views and monotone backgrounds. Visit our project page at: https://umautobots.github.io/lihi_gs

In the rapidly evolving field of Large Language Models (LLMs), ensuring safety is a crucial and widely discussed topic. However, existing works often overlook the geo-diversity of cultural and legal standards across the world. To demonstrate the challenges posed by geo-diverse safety standards, we introduce SafeWorld, a novel benchmark specifically designed to evaluate LLMs' ability to generate responses that are not only helpful but also culturally sensitive and legally compliant across diverse global contexts. SafeWorld encompasses 2,342 test user queries, each grounded in high-quality, human-verified cultural norms and legal policies from 50 countries and 493 regions/races. On top of it, we propose a multi-dimensional automatic safety evaluation framework that assesses the contextual appropriateness, accuracy, and comprehensiveness of responses. Our evaluations reveal that current LLMs struggle to meet these criteria. To enhance LLMs' alignment with geo-diverse safety standards, we synthesize helpful preference pairs for Direct Preference Optimization (DPO) alignment training. The preference pair construction aims to encourage LLMs to behave appropriately and provide precise references to relevant cultural norms and policies when necessary. Our trained SafeWorldLM outperforms all competing models, including GPT-4o on all three evaluation dimensions by a large margin. Global human evaluators also note a nearly 20% higher winning rate in helpfulness and harmfulness evaluation. Our code and data can be found here: https://github.com/PlusLabNLP/SafeWorld., Comment: Accepted by NeurIPS 2024

This project investigates factors that influence the perceived helpfulness of Amazon product reviews through machine learning techniques. After extensive feature analysis and correlation testing, we identified key metadata characteristics that serve as strong predictors of review helpfulness. While we initially explored natural language processing approaches using TextBlob for sentiment analysis, our final model focuses on metadata features that demonstrated more significant correlations, including the number of images per review, reviewer's historical helpful votes, and temporal aspects of the review. The data pipeline encompasses careful preprocessing and feature standardization steps to prepare the input for model training. Through systematic evaluation of different feature combinations, we discovered that metadata elements we choose using a threshold provide reliable signals when combined for predicting how helpful other Amazon users will find a review. This insight suggests that contextual and user-behavioral factors may be more indicative of review helpfulness than the linguistic content itself.

As events progress, news articles often update with new information: if we are not cautious, we risk propagating outdated facts. In this work, we hypothesize that linguistic features indicate factual fluidity, and that we can predict which facts in a news article will update using solely the text of a news article (i.e. not external resources like search engines). We test this hypothesis, first, by isolating fact-updates in large news revisions corpora. News articles may update for many reasons (e.g. factual, stylistic, narrative). We introduce the NewsEdits 2.0 taxonomy, an edit-intentions schema that separates fact updates from stylistic and narrative updates in news writing. We annotate over 9,200 pairs of sentence revisions and train high-scoring ensemble models to apply this schema. Then, taking a large dataset of silver-labeled pairs, we show that we can predict when facts will update in older article drafts with high precision. Finally, to demonstrate the usefulness of these findings, we construct a language model question asking (LLM-QA) abstention task. We wish the LLM to abstain from answering questions when information is likely to become outdated. Using our predictions, we show, LLM absention reaches near oracle levels of accuracy., Comment: 9 pages main body, 11 pages appendix

We study the quantitative pointwise behavior of solutions to the Boltzmann equation for hard potentials and Maxwellian molecules, which generalize the hard sphere case introduced by Liu-Yu in 2004 (Comm. Pure Appl. Math. 57:1543-1608, 2004). The large time behavior of the solution is dominated by fluid structures, similar to the hard sphere case. However, unlike hard sphere, the spatial decay here depends on the potential power $\gamma$ and the initial velocity weight. A key challenge in this problem is the loss of velocity weight in linear estimates, which makes standard nonlinear iteration infeasible. To address this, we develop an Enhanced Mixture Lemma, demonstrating that mixing the transport and gain parts of the linearized collision operator can generate arbitrary-order regularity and decay in both space and velocity variables. This allows us to decompose the linearized solution into fluid (arbitrary regularity and velocity decay) and particle (rapid space-time decay, but with loss of velocity decay) parts, making it possible to solve the nonlinear problem through this particle-fluid duality., Comment: 37 pages

We present GazeGen, a user interaction system that generates visual content (images and videos) for locations indicated by the user's eye gaze. GazeGen allows intuitive manipulation of visual content by targeting regions of interest with gaze. Using advanced techniques in object detection and generative AI, GazeGen performs gaze-controlled image adding/deleting, repositioning, and surface style changes of image objects, and converts static images into videos. Central to GazeGen is the DFT Gaze (Distilled and Fine-Tuned Gaze) agent, an ultra-lightweight model with only 281K parameters, performing accurate real-time gaze predictions tailored to individual users' eyes on small edge devices. GazeGen is the first system to combine visual content generation with real-time gaze estimation, made possible exclusively by DFT Gaze. This real-time gaze estimation enables various visual content generation tasks, all controlled by the user's gaze. The input for DFT Gaze is the user's eye images, while the inputs for visual content generation are the user's view and the predicted gaze point from DFT Gaze. To achieve efficient gaze predictions, we derive the small model from a large model (10x larger) via novel knowledge distillation and personal adaptation techniques. We integrate knowledge distillation with a masked autoencoder, developing a compact yet powerful gaze estimation model. This model is further fine-tuned with Adapters, enabling highly accurate and personalized gaze predictions with minimal user input. DFT Gaze ensures low-latency and precise gaze tracking, supporting a wide range of gaze-driven tasks. We validate the performance of DFT Gaze on AEA and OpenEDS2020 benchmarks, demonstrating low angular gaze error and low latency on the edge device (Raspberry Pi 4). Furthermore, we describe applications of GazeGen, illustrating its versatility and effectiveness in various usage scenarios., Comment: 12 pages, 10 figures

Interoperability is the key to implementing a wide range of energy systems applications. It involves the seamless cooperation of different methods and components. With smart energy systems, interoperability faces challenges due to integrating differ-ent approaches and technologies. This includes dealing with heterogeneous approaches with various communication proto-cols and data formats. However, it is essential for smart energy systems to carry out thorough interoperability tests. They are usually diverse, and challenging, thus requiring careful consideration of compatibility issues and complex integration scenari-os. Overcoming these challenges requires a systematic approach that includes thorough test planning, rigorous testing, and continuous test monitoring. Although numerous testing approaches exist, most are more developed at the component/device level than at the system level. Consequently, there are few approaches and related facilities to test the interoperability of smart energy approaches and solutions at the system level. This work analyses existing interoperability test concepts, identi-fies enablers and the potential for harmonisation of procedures, and proposes further developments of these approaches., Comment: 14th Mediterranean Conference on Power Generation Transmission, Distribution and Energy Conversion (MED POWER 2024)

Customer Relationship Management (CRM) systems are vital for modern enterprises, providing a foundation for managing customer interactions and data. Integrating AI agents into CRM systems can automate routine processes and enhance personalized service. However, deploying and evaluating these agents is challenging due to the lack of realistic benchmarks that reflect the complexity of real-world CRM tasks. To address this issue, we introduce CRMArena, a novel benchmark designed to evaluate AI agents on realistic tasks grounded in professional work environments. Following guidance from CRM experts and industry best practices, we designed CRMArena with nine customer service tasks distributed across three personas: service agent, analyst, and manager. The benchmark includes 16 commonly used industrial objects (e.g., account, order, knowledge article, case) with high interconnectivity, along with latent variables (e.g., complaint habits, policy violations) to simulate realistic data distributions. Experimental results reveal that state-of-the-art LLM agents succeed in less than 40% of the tasks with ReAct prompting, and less than 55% even with function-calling abilities. Our findings highlight the need for enhanced agent capabilities in function-calling and rule-following to be deployed in real-world work environments. CRMArena is an open challenge to the community: systems that can reliably complete tasks showcase direct business value in a popular work environment.

Current face anonymization techniques often depend on identity loss calculated by face recognition models, which can be inaccurate and unreliable. Additionally, many methods require supplementary data such as facial landmarks and masks to guide the synthesis process. In contrast, our approach uses diffusion models with only a reconstruction loss, eliminating the need for facial landmarks or masks while still producing images with intricate, fine-grained details. We validated our results on two public benchmarks through both quantitative and qualitative evaluations. Our model achieves state-of-the-art performance in three key areas: identity anonymization, facial attribute preservation, and image quality. Beyond its primary function of anonymization, our model can also perform face swapping tasks by incorporating an additional facial image as input, demonstrating its versatility and potential for diverse applications. Our code and models are available at https://github.com/hanweikung/face_anon_simple .

As large language models (LLMs) expand into performing as agents for real-world applications beyond traditional NLP tasks, evaluating their robustness becomes increasingly important. However, existing benchmarks often overlook critical dimensions like cultural and social awareness. To address these, we introduce CASA, a benchmark designed to assess LLM agents' sensitivity to cultural and social norms across two web-based tasks: online shopping and social discussion forums. Our approach evaluates LLM agents' ability to detect and appropriately respond to norm-violating user queries and observations. Furthermore, we propose a comprehensive evaluation framework that measures awareness coverage, helpfulness in managing user queries, and the violation rate when facing misleading web content. Experiments show that current LLMs perform significantly better in non-agent than in web-based agent environments, with agents achieving less than 10% awareness coverage and over 40% violation rates. To improve performance, we explore two methods: prompting and fine-tuning, and find that combining both methods can offer complementary advantages -- fine-tuning on culture-specific datasets significantly enhances the agents' ability to generalize across different regions, while prompting boosts the agents' ability to navigate complex tasks. These findings highlight the importance of constantly benchmarking LLM agents' cultural and social awareness during the development cycle., Comment: Work in progress

We introduce a novel approach to reduce the number of times required for reprogramming memristors on bit-sliced compute-in-memory crossbars for deep neural networks (DNNs). Our idea addresses the limited non-volatile memory endurance, which restrict the number of times they can be reprogrammed. To reduce reprogramming demands, we employ two techniques: (1) we organize weights into sorted sections to schedule reprogramming of similar crossbars, maximizing memristor state reuse, and (2) we reprogram only a fraction of randomly selected memristors in low-order columns, leveraging their bit-level distribution and recognizing their relatively small impact on model accuracy. We evaluate our approach for state-of-the-art models on the ImageNet-1K dataset. We demonstrate a substantial reduction in crossbar reprogramming by 3.7x for ResNet-50 and 21x for ViT-Base, while maintaining model accuracy within a 1% margin., Comment: 5 pages, 10 figures

Coupling black hole (BH) feeding and feedback involves interactions across vast spatial and temporal scales that is computationally challenging. Tracking gas inflows and outflows from kilo-parsec scales to the event horizon for non-spinning BHs in the presence of strong magnetic fields, Cho et al. (2023, 2024) report strong suppression of accretion on horizon scales and low (2%) feedback efficiency. In this letter, we explore the impact of these findings for the supermassive BHs M87* and Sgr A*, using high-resolution, non-cosmological, magnetohydrodynamic (MHD) simulations with the Feedback In Realistic Environments (FIRE-2) model. With no feedback, we find rapid BH growth due to "cooling flows," and for 2% efficiency feedback, while accretion is suppressed, the rates still remain higher than constraints from Event Horizon Telescope (EHT) data (Event Horizon Telescope Collaboration et al. 2021, 2022) for M87* and Sgr A*. To match EHT observations of M87*, a feedback efficiency greater than 15% is required, suggesting the need to include enhanced feedback from BH spin. Similarly, a feedback efficiency of $>15\%$ is needed for Sgr A* to match the estimated observed star formation rate of $\lesssim 2 {\rm M_\odot}$ yr$^{-1}$. However, even with 100% feedback efficiency, the accretion rate onto Sgr A* matches with EHT data only on rare occasions in the simulations, suggesting that Sgr A* is likely in a temporary quiescent phase currently. Bridging accretion and feedback across scales, we conclude that higher feedback efficiency, possibly due to non-zero BH spin, is necessary to suppress "cooling flows" and match observed accretion and star formation rates in M87* and Sgr A*., Comment: 13 pages, 4 figures

We introduce $\textit{sorted weight sectioning}$ (SWS): a weight allocation algorithm that places sorted deep neural network (DNN) weight sections on bit-sliced compute-in-memory (CIM) crossbars to reduce analog-to-digital converter (ADC) energy consumption. Data conversions are the most energy-intensive process in crossbar operation. SWS effectively reduces this cost leveraging (1) small weights and (2) zero weights (weight sparsity). DNN weights follow bell-shaped distributions, with most weights near zero. Using SWS, we only need low-order crossbar columns for sections with low-magnitude weights. This reduces the quantity and resolution of ADCs used, exponentially decreasing ADC energy costs without significantly degrading DNN accuracy. Unstructured sparsification further sharpens the weight distribution with small accuracy loss. However, it presents challenges in hardware tracking of zeros: we cannot switch zero rows to other layer weights in unsorted crossbars without index matching. SWS efficiently addresses unstructured sparse models using offline remapping of zeros into earlier sections, which reveals full sparsity potential and maximizes energy efficiency. Our method reduces ADC energy use by 89.5% on unstructured sparse BERT models. Overall, this paper introduces a novel algorithm to promote energy-efficient CIM crossbars for unstructured sparse DNN workloads., Comment: 5 pages, 4 figures

In an effort to understand the presence of super-virial gas detected in the Milky Way, we present our findings from isolated galaxy simulations of Milky Way-like systems using GIZMO with the FIRE-2 (Feedback In Realistic Environments) stellar feedback model. It unveils the presence of a significant super-virial temperature ( $T>6\times10^6$K) gas component within 20 kpc from the galactic center. This super-virial gas has a mass of $1-2\times10^7$ ${\rm M_\odot}$ and is found close to the disk, where typical gas densities are $0.004-0.01 \, \rm cm^{-3}$. We find that some of the virial gas ($T\sim10^6$K) forms a rotating hot inflow, where gravitational energy is converted to heat mainly via compressive heating. This process causes gas infalling close to the rotation axis to reach super-virial temperatures just before cooling and joining the disk. Stellar feedback heating accounts for less than 1% of the super-virial gas, indicating its minimal influence despite expectations. Even in scenarios with no stellar feedback effects considered, abundant super-virial gas persists, highlighting the dominance of alternative heating mechanisms. We also show that cosmic rays do not have a significant effect on heating the gas to a super-virial temperature. Our study illuminates the intricate dynamics of hot virial and super-virial gas surrounding Milky Way-like galaxies, emphasizing the prominent role of infall-driven and compressive heating processes in shaping thermal evolution., Comment: Submitted to ApJ, 14 pages, 8 figures, 1 table

We consider the H\"older regularity of solutions to the steady Boltzmann equation with in-flow boundary condition in bounded and strictly convex domains $\Omega\subset\mathbb{R}^{3}$ for gases with cutoff soft potential $(-3<\gamma<0)$. We prove that there is a unique solution with a bounded $L^{\infty}$ norm in space and velocity. This solution is H\"older continuous, and it's order depends not only on the regularity of the incoming boundary data, but also on the potential power $\gamma$. The result for modulated soft potential case $-2<\gamma<0$ is similar to hard potential case $(0\leq\gamma<1)$ since we have $C^{1}$ velocity regularity from collision part. However, we observe that for very soft potential case $(-3<\gamma\leq -2)$, the regularity in velocity obtained by the collision part is lower (H\"older only), but the boundary regularity still can transfer to solution (in both space and velocity) by transport and collision part under the restriction of $\gamma$.

The early growth of black holes (BHs) in atomic-cooling halos is likely influenced by feedback on the surrounding gas. While the effects of radiative feedback are well-documented, mechanical feedback, particularly from AGN jets, has been comparatively less explored. Building on our previous work that examined the growth of a 100 ${M_\odot}$ BH in a constant density environment regulated by AGN jets, we expand the initial BH mass range from 1 to $10^4$ ${M_\odot}$ and adopt a more realistic density profile for atomic-cooling halos. We reaffirm the validity of our analytic models for jet cocoon propagation and feedback regulation. We identify several critical radii-namely, the terminal radius of jet cocoon propagation, the isotropization radius of the jet cocoon, and the core radius of the atomic-cooling halo-that are crucial in determining BH growth given specific gas properties and jet feedback parameters. In a significant portion of the parameter space, our findings show that jet feedback substantially disrupts the halo's core during the initial feedback episode, preventing BH growth beyond $10^4$ ${M_\odot}$. Conversely, conditions characterized by low jet velocities and high gas densities enable sustained BH growth over extended periods. We provide a prediction for the black hole mass growth as a function of time and feedback parameters. We found that, to form a supermassive BH ($>10^6 {M_\odot}$) within 1 Gyr entirely by accreting gas from an atomic-cooling halo, the jet energy feedback efficiency must be $\lesssim 10^{-4} \dot{M}_{BH} c^2$ even if the seed BH mass is $10^4 {M_\odot}$., Comment: 20 pages, 14 figures; submitted to MNRAS; comments welcome!

BACKGROUND: Kidney transplant is the treatment of choice for patients with end-stage renal disease. Early detection of allograft injury is important to delay or prevent irreversible damage. PURPOSE: To investigate the feasibility of hyperpolarized (HP) [1-13C]pyruvate MRI for assessing kidney allograft metabolism. SUBJECTS: 6 participants (mean age, 45.2 +- 12.4 years, 2 females) scheduled for kidney allograft biopsy and 5 patients (mean age, 59.6 +- 10.4 years, 2 females) with renal cell carcinoma (RCC). ASSESSMENT: Five of the six kidney allograft participants underwent biopsy after MRI. Estimated glomerular filtration rate (eGFR) and urine protein-to-creatine ratio (uPCR) were collected within 4 weeks of MRI. Kidney metabolism was quantified from HP [1-13C]pyruvate MRI using the lactate-to-pyruvate ratio in allograft kidneys and non-tumor bearing kidneys from RCC patients. RESULTS: Biopsy was performed a mean of 9 days (range 5-19 days) after HP [1-13C]pyruvate MRI. Three biopsies were normal, one showed low-grade fibrosis and one showed moderate microvascular inflammation. All had stable functioning allografts with eGFR > 60 mL/min/1.73 m2 and normal uPCR. One participant who did not undergo biopsy had reduced eGFR of 49 mL/min/1.73 m2 and elevated uPCR. The mean lactate-to-pyruvate ratio was 0.373 in participants with normal findings (n = 3) and 0.552 in participants with abnormal findings (n = 2). The lactate-to-pyruvate ratio was highest (0.847) in the participant with reduced eGFR and elevated uPRC. Native non-tumor bearing kidneys had a mean lactate-to-pyruvate ratio of 0.309. DATA CONCLUSION: Stable allografts with normal findings at biopsy showed lactate-to-pyruvate ratios similar to native non-tumor bearing kidneys, whereas allografts with abnormal findings showed higher lactate-to-pyruvate ratios., Comment: J Magn Reson Imaging (2024)

Scene coordinate regression (SCR) methods have emerged as a promising area of research due to their potential for accurate visual localization. However, many existing SCR approaches train on samples from all image regions, including dynamic objects and texture-less areas. Utilizing these areas for optimization during training can potentially hamper the overall performance and efficiency of the model. In this study, we first perform an in-depth analysis to validate the adverse impacts of these areas. Drawing inspiration from our analysis, we then introduce an error-guided feature selection (EGFS) mechanism, in tandem with the use of the Segment Anything Model (SAM). This mechanism seeds low reprojection areas as prompts and expands them into error-guided masks, and then utilizes these masks to sample points and filter out problematic areas in an iterative manner. The experiments demonstrate that our method outperforms existing SCR approaches that do not rely on 3D information on the Cambridge Landmarks and Indoor6 datasets., Comment: ECCV2024

To overcome the burden on the memory size and bandwidth due to ever-increasing size of large language models (LLMs), aggressive weight quantization has been recently studied, while lacking research on quantizing activations. In this paper, we present a hardware-software co-design method that results in an energy-efficient LLM accelerator, named OPAL, for generation tasks. First of all, a novel activation quantization method that leverages the microscaling data format while preserving several outliers per sub-tensor block (e.g., four out of 128 elements) is proposed. Second, on top of preserving outliers, mixed precision is utilized that sets 5-bit for inputs to sensitive layers in the decoder block of an LLM, while keeping inputs to less sensitive layers to 3-bit. Finally, we present the OPAL hardware architecture that consists of FP units for handling outliers and vectorized INT multipliers for dominant non-outlier related operations. In addition, OPAL uses log2-based approximation on softmax operations that only requires shift and subtraction to maximize power efficiency. As a result, we are able to improve the energy efficiency by 1.6~2.2x, and reduce the area by 2.4~3.1x with negligible accuracy loss, i.e., <1 perplexity increase., Comment: 7 pages, 8 figures, DAC2024 accepted

Trust is a subjective yet fundamental component of human-computer interaction, and is a determining factor in shaping the efficacy of data visualizations. Prior research has identified five dimensions of trust assessment in visualizations (credibility, clarity, reliability, familiarity, and confidence), and observed that these dimensions tend to vary predictably along with certain features of the visualization being evaluated. This raises a further question: how do the design features driving viewers trust assessment vary with the characteristics of the viewers themselves? By reanalyzing data from these studies through the lens of individual differences, we build a more detailed map of the relationships between design features, individual characteristics, and trust behaviors. In particular, we model the distinct contributions of endogenous design features (such as visualization type, or the use of color) and exogenous user characteristics (such as visualization literacy), as well as the interactions between them. We then use these findings to make recommendations for individualized and adaptive visualization design.

We study the Boltzmann equation with hard sphere in a near-equilibrium setting. The initial data is compactly supported in the space variable and has a polynomial tail in the microscopic velocity. We show that the solution can be decomposed into a particle-like part (polynomial tail) and a fluid-like part (Gaussian tail). The particle-like part decays exponentially in both space and time, while the fluid-like part corresponds to the behavior of the compressible Navier-Stokes equation, which dominates the long time behavior and exhibits rich wave motion. The nonlinear wave interactions in the fluid-like part are precisely characterized and therefore we are able to distinguish the linear and nonlinear wave of the solution. It is notable that although the solution has polynomial tail in the velocity initially, the transition process from the polynomial to the Gaussian tail can be quantitatively revealed due to the collision with the background global Maxwellian.

Generative models have surged in popularity recently due to their ability to produce high-quality images and video. However, steering these models to produce images with specific attributes and precise control remains challenging. Humans, particularly their faces, are central to content generation due to their ability to convey rich expressions and intent. Current generative models mostly generate flat neutral expressions and characterless smiles without authenticity. Other basic expressions like anger are possible, but are limited to the stereotypical expression, while other unconventional facial expressions like doubtful are difficult to reliably generate. In this work, we propose the use of AUs (action units) for facial expression control in face generation. AUs describe individual facial muscle movements based on facial anatomy, allowing precise and localized control over the intensity of facial movements. By combining different action units, we unlock the ability to create unconventional facial expressions that go beyond typical emotional models, enabling nuanced and authentic reactions reflective of real-world expressions. The proposed method can be seamlessly integrated with both text and image prompts using adapters, offering precise and intuitive control of the generated results. Code and dataset are available in {https://github.com/tvaranka/fineface}.

In the past two decades, various classes of topological materials have been discovered, spanning topological insulators, semimetals, and metals. While the observation and understanding of the topology of a material has been a primary focus so far, the precise and easy control of topology in a single material remains largely unexplored. Here, we demonstrate full experimental control over the topological Dirac nodal loop in the square-net material LaSb$_\mathrm{x}$Te$_\mathrm{2-x}$ by chemical substitution and electron doping. Using angle-resolved photoemission spectroscopy (ARPES), we show that changing the antimony concentration x from 0.9 to 1.0 in the bulk opens a gap as large as 400 meV in the nodal loop. Our symmetry analysis based on single-crystal X-ray diffraction and a minimal tight binding model establishes that the breaking of \textit{n} glide symmetry in the square-net layer is responsible for the opening of the gap. Remarkably, we can also realize this topological phase transition \textit{in situ} on the surface of LaSb$_\mathrm{x}$Te$_\mathrm{2-x}$ by chemical gating using potassium deposition, which enables the reversible switching of the topology from gapped to gapless nodal loop. The underlying control parameter for the structural and topological transition in the bulk and on the surface is the electron concentration. It opens a pathway towards applications in devices based on switching topology by electrostatic gating.

We undertake a comprehensive investigation into the distribution of insitu stars within Milky Way-like galaxies, leveraging TNG50 simulations and comparing their predictions with data from the H3 survey. Our analysis reveals that 28% of galaxies demonstrate reasonable agreement with H3, while only 12% exhibit excellent alignment in their profiles, regardless of the specific spatial cut employed to define insitu stars. To uncover the underlying factors contributing to deviations between TNG50 and H3 distributions, we scrutinize correlation coefficients among internal drivers(e.g., virial radius, star formation rate [SFR]) and merger-related parameters (such as the effective mass-ratio, mean distance, average redshift, total number of mergers, average spin-ratio and maximum spin alignment between merging galaxies). Notably, we identify significant correlations between deviations from observational data and key parameters such as the median slope of virial radius, mean SFR values, and the rate of SFR change across different redshift scans. Furthermore, positive correlations emerge between deviations from observational data and parameters related to galaxy mergers. We validate these correlations using the Random Forest Regression method. Our findings underscore the invaluable insights provided by the H3 survey in unraveling the cosmic history of galaxies akin to the Milky Way, thereby advancing our understanding of galactic evolution and shedding light on the formation and evolution of Milky Way-like galaxies in cosmological simulations., Comment: 20 pages, 13 figures

We introduce our method, conformal highest conditional density sets (CHCDS), that forms conformal prediction sets using existing estimated conditional highest density predictive regions. We prove the validity of the method and that conformal adjustment is negligible under some regularity conditions. In particular, if we correctly specify the underlying conditional density estimator, the conformal adjustment will be negligible. When the underlying model is incorrect, the conformal adjustment provides guaranteed nominal unconditional coverage. We compare the proposed method via simulation and a real data analysis to other existing methods. Our numerical results show that the flexibility of being able to use any existing conditional density estimation method is a large advantage for CHCDS compared to existing methods.

Despite the success of Large Language Models (LLMs) on various tasks following human instructions, controlling model generation at inference time poses a persistent challenge. In this paper, we introduce Ctrl-G, an adaptable framework that facilitates tractable and flexible control of LLM generation to reliably follow logical constraints. Ctrl-G combines any production-ready LLM with a Hidden Markov Model, enabling LLM outputs to adhere to logical constraints represented as deterministic finite automata. We show that Ctrl-G, when applied to a TULU2-7B model, outperforms GPT3.5 and GPT4 on the task of interactive text editing: specifically, for the task of generating text insertions/continuations following logical constraints, Ctrl-G achieves over 30% higher satisfaction rate in human evaluation compared to GPT4. When applied to medium-size language models (e.g., GPT2-large), Ctrl-G also beats its counterparts for constrained generation by large margins on standard benchmarks. Additionally, as a proof-of-concept study, we experiment Ctrl-G on the Grade School Math benchmark to assist LLM reasoning, foreshadowing the application of Ctrl-G, as well as other constrained generation approaches, beyond traditional language generation tasks.

Sustainable Development Goals (SDGs) give the UN a road map for development with Agenda 2030 as a target. SDG3 "Good Health and Well-Being" ensures healthy lives and promotes well-being for all ages. Digital technologies can support SDG3. Burnout and even depression could be reduced by encouraging better preventive health. Due to the lack of patient knowledge and focus to take care of their health, it is necessary to help patients before it is too late. New trends such as positive psychology and mindfulness are highly encouraged in the USA. Digital Twins (DTs) can help with the continuous monitoring of emotion using physiological signals (e.g., collected via wearables). DTs facilitate monitoring and provide constant health insight to improve quality of life and well-being with better personalization. Healthcare DTs challenges are standardizing data formats, communication protocols, and data exchange mechanisms. As an example, ISO has the ISO/IEC JTC 1/SC 41 Internet of Things (IoT) and DTs Working Group, with standards such as "ISO/IEC 21823-3:2021 IoT - Interoperability for IoT Systems - Part 3 Semantic interoperability", "ISO/IEC CD 30178 - IoT - Data format, value and coding". To achieve those data integration and knowledge challenges, we designed the Mental Health Knowledge Graph (ontology and dataset) to boost mental health. As an example, explicit knowledge is described such as chocolate contains magnesium which is recommended for depression. The Knowledge Graph (KG) acquires knowledge from ontology-based mental health projects classified within the LOV4IoT ontology catalog (Emotion, Depression, and Mental Health). Furthermore, the KG is mapped to standards when possible. Standards from ETSI SmartM2M can be used such as SAREF4EHAW to represent medical devices and sensors, but also ITU/WHO, ISO, W3C, NIST, and IEEE standards relevant to mental health can be considered., Comment: 20 pages, Book chapter, Smart Technologies for Achieving Good Health and Well-Being: Towards Sustainable Development Goal, Taylor & Francis

We propose a new method for finding the highest predictive density set or region using signed conformal inference. The proposed method is computationally efficient, while also carrying conformal coverage guarantees. We prove that under, mild regularity conditions, the conformal prediction set is asymptotically close to its oracle counterpart. The efficacy of the method is illustrated through simulations and real applications.

We propose conformal hyperrectangular prediction regions for multi-target regression. We propose split conformal prediction algorithms for both point and quantile regression to form hyperrectangular prediction regions, which allow for easy marginal interpretation and do not require covariance estimation. In practice, it is preferable that a prediction region is balanced, that is, having identical marginal prediction coverage, since prediction accuracy is generally equally important across components of the response vector. The proposed algorithms possess two desirable properties, namely, tight asymptotic overall nominal coverage as well as asymptotic balance, that is, identical asymptotic marginal coverage, under mild conditions. We then compare our methods to some existing methods on both simulated and real data sets. Our simulation results and real data analysis show that our methods outperform existing methods while achieving the desired nominal coverage and good balance between dimensions.

Simulating accretion and feedback from the horizon scale of supermassive black holes (SMBHs) out to galactic scales is challenging because of the vast range of scales involved. Elaborating on \citet{Cho2023}, we describe and test a ``multi-zone'' technique which is designed to tackle this difficult problem in 3D general relativistic magnetohydrodynamic (GRMHD) simulations. While short-timescale variability should be interpreted with caution, the method is demonstrated to be well-suited for finding dynamical steady-states over a wide dynamic range. We simulate accretion on a non-spinning SMBH ($a_*=0$) using initial conditions and the external galactic potential from a large scale galaxy simulation, and achieve steady state over 8 decades in radius. As found in \citet{Cho2023}, the density scales with radius as $\rho \propto r^{-1}$ inside the Bondi radius $R_B$, which is located at $R_B=2\times 10^5 \,r_g$ ($\approx 60\,{\rm pc}$ for M87) where $r_g$ is the gravitational radius of the SMBH; the plasma-$\beta\sim$ unity, indicating an extended magnetically arrested state; the mass accretion rate $\dot{M}$ is $\approx 1\%$ of the analytical Bondi accretion rate $\dot{M}_B$; and there is continuous energy feedback out to $\approx 100R_B$ (or beyond $>\,{\rm kpc}$) at a rate $\approx 0.02 \dot{M}c^2$. Surprisingly, no ordered rotation in the external medium survives as the magnetized gas flows to smaller radii, and the final steady solution is very similar to when the exterior has no rotation. Using the multi-zone method, we simulate GRMHD accretion over a wide range of Bondi radii, $R_{\rm B} \sim 10^2 - 10^7\,r_{\rm g}$, and find that $\dot{M}/\dot{M}_B\approx (R_B/6\, r_g)^{-0.5}$., Comment: 21 pages, 9 figures, submitted to ApJ

This paper investigates the development and evaluation of machine translation models from Cantonese to English, where we propose a novel approach to tackle low-resource language translations. The main objectives of the study are to develop a model that can effectively translate Cantonese to English and evaluate it against state-of-the-art commercial models. To achieve this, a new parallel corpus has been created by combining different available corpora online with preprocessing and cleaning. In addition, a monolingual Cantonese dataset has been created through web scraping to aid the synthetic parallel corpus generation. Following the data collection process, several approaches, including fine-tuning models, back-translation, and model switch, have been used. The translation quality of models has been evaluated with multiple quality metrics, including lexicon-based metrics (SacreBLEU and hLEPOR) and embedding-space metrics (COMET and BERTscore). Based on the automatic metrics, the best model is selected and compared against the 2 best commercial translators using the human evaluation framework HOPES. The best model proposed in this investigation (NLLB-mBART) with model switch mechanisms has reached comparable and even better automatic evaluation scores against State-of-the-art commercial models (Bing and Baidu Translators), with a SacreBLEU score of 16.8 on our test set. Furthermore, an open-source web application has been developed to allow users to translate between Cantonese and English, with the different trained models available for effective comparisons between models from this investigation and users. CANTONMT is available at https://github.com/kenrickkung/CantoneseTranslation, Comment: on-going work, 30 pages