Search

Multimodal foundation models, such as Gemini and ChatGPT, have revolutionized human-machine interactions by seamlessly integrating various forms of data. Developing a universal spoken language model that comprehends a wide range of natural language instructions is critical for bridging communication gaps and facilitating more intuitive interactions. However, the absence of a comprehensive evaluation benchmark poses a significant challenge. We present Dynamic-SUPERB Phase-2, an open and evolving benchmark for the comprehensive evaluation of instruction-based universal speech models. Building upon the first generation, this second version incorporates 125 new tasks contributed collaboratively by the global research community, expanding the benchmark to a total of 180 tasks, making it the largest benchmark for speech and audio evaluation. While the first generation of Dynamic-SUPERB was limited to classification tasks, Dynamic-SUPERB Phase-2 broadens its evaluation capabilities by introducing a wide array of novel and diverse tasks, including regression and sequence generation, across speech, music, and environmental audio. Evaluation results indicate that none of the models performed well universally. SALMONN-13B excelled in English ASR, while WavLLM demonstrated high accuracy in emotion recognition, but current models still require further innovations to handle a broader range of tasks. We will soon open-source all task data and the evaluation pipeline.

Realistic hair motion is crucial for high-quality avatars, but it is often limited by the computational resources available for real-time applications. To address this challenge, we propose a novel neural approach to predict physically plausible hair deformations that generalizes to various body poses, shapes, and hairstyles. Our model is trained using a self-supervised loss, eliminating the need for expensive data generation and storage. We demonstrate our method's effectiveness through numerous results across a wide range of pose and shape variations, showcasing its robust generalization capabilities and temporally smooth results. Our approach is highly suitable for real-time applications with an inference time of only a few milliseconds on consumer hardware and its ability to scale to predicting the drape of 1000 grooms in 0.3 seconds.

We present an analysis of three near-infrared (NIR; 1.0-2.4 $\mu$m) spectra of the SN 2003fg-like/"super-Chandrasekhar" type Ia supernovae (SNe Ia) SN 2009dc, SN 2020hvf, and SN 2022pul at respective phases +372, +296, and +294~d relative to the epoch of $B$-band maximum. We find that all objects in our sample have asymmetric, or "tilted", [Fe~II] 1.257 and 1.644 $\mu$m profiles. We quantify the asymmetry of these features using five methods: velocity at peak flux, profile tilts, residual testing, velocity fitting, and comparison to deflagration-detonation transition models. Our results demonstrate that, while the profiles of the [Fe II] 1.257 and 1.644 $\mu$m features are widely varied between 2003fg-likes, these features are correlated in shape within the same SN. This implies that line blending is most likely not the dominant cause of the asymmetries inferred from these profiles. Instead, it is more plausible that 2003fg-like SNe have aspherical chemical distributions in their inner regions. These distributions may come from aspherical progenitor systems, such as double white dwarf mergers, or off-center delayed-detonation explosions of Chandrasekhar-mass Carbon-Oxygen white dwarfs. Additional late-phase NIR observation of 2003fg-like SNe and detailed 3-D NLTE modeling of these two explosion scenarios are encouraged., Comment: Submitted to ApJ

While recent works on blind face image restoration have successfully produced impressive high-quality (HQ) images with abundant details from low-quality (LQ) input images, the generated content may not accurately reflect the real appearance of a person. To address this problem, incorporating well-shot personal images as additional reference inputs could be a promising strategy. Inspired by the recent success of the Latent Diffusion Model (LDM), we propose ReF-LDM, an adaptation of LDM designed to generate HQ face images conditioned on one LQ image and multiple HQ reference images. Our model integrates an effective and efficient mechanism, CacheKV, to leverage the reference images during the generation process. Additionally, we design a timestep-scaled identity loss, enabling our LDM-based model to focus on learning the discriminating features of human faces. Lastly, we construct FFHQ-Ref, a dataset consisting of 20,405 high-quality (HQ) face images with corresponding reference images, which can serve as both training and evaluation data for reference-based face restoration models., Comment: NeurIPS 2024, project page https://chiweihsiao.github.io/refldm.github.io/

We have obtained near-infrared ($0.80-2.45\mu$m) spectra of the recurrent nova LMCN 1968-12a on two occasions during its 2024 August eruption. This is the first near-infrared spectroscopy of an extragalactic nova. The initial spectrum, on day 8.48, caught the nova in the coronal phase, with the [SiX] $1.43\mu$m line being extremely strong. This line had a luminosity of $\sim95$L$_\odot$, and is clearly a very powerful coolant. Its presence, together with the absence of [SiIX] 1.56$\mu$m, implies a coronal temperature $\gtrsim3\times10^6$K, possibly amongst the highest recorded coronal temperature in a nova eruption. With the exception of the [SiX] line, the near-infrared spectra are remarkable for being devoid of metal lines. We suggest that this is due, in part, to the exceptionally high temperature of the coronal gas, causing ions, whose emission lines would normally appear in the near-infrared spectrum, to be collisionally ionised to higher stages., Comment: Accepted for publication in MNRAS, 4 December 2024

The BL Lac object 3C 371 is one of the targets that are regularly monitored by the Whole Earth Blazar Telescope (WEBT) Collaboration to study blazar variability on both short and long timescales. We aim to evaluate the long-term multiwavelength (MWL) behaviour of 3C 371, comparing it with the results derived for its optical emission in our previous study. For this, we make use of the multi-band campaigns organized by the WEBT Collaboration in optical and radio between January 2018 and December 2020, and of public data from Swift and Fermi satellites and the MOJAVE Very Large Interferometry programme. We evaluate the variability shown by the source in each band with the amplitude variability quantification, as well as possible interband correlation using the z-Discrete Correlation Function. We also present a deep analysis of the optical-UV, X-ray and $\gamma$-ray spectral variability. With the MOJAVE data we perform a kinematics analysis, looking for components propagating along the jet, calculating its kinematics parameters. This set of parameters is later used for the interpretation of the source MWL behaviour, modelling the broadband spectral energy distribution (SED) of the source with theoretical blazar emission scenarios., Comment: Accepted in A&A, 23 pages, 15 figures

We investigate the sextet $b$-baryon decay processes $\Omega_b\to J/\psi\Omega^{(*)}$, where $\Omega^*$ represents the $1P$-, $1D$- and $2S$-wave excited $\Omega$ hyperons in the spectroscopy. Using the constituent quark model, we obtain ${\cal B}(\Omega_b \to J/\psi\Omega)=2.1\times 10^{-4}$, which agrees with the previous studies to the order of magnitude. By identifying $\Omega(2012)$ as $\Omega(1^2P_{3/2^-})$, ${\cal B}(\Omega_b \to J/\psi\Omega(2012))=2.7\times 10^{-4}$ can be similarly significant. Additionally, $\Omega(1^2D_{5/2^+})$ and $\Omega(1^4D_{3/2^+},1^4D_{5/2^+})$ states exhibit production rates of 0.6, and (0.7, 1.0), respectively, relative to their ground-state counterpart. Notably, our findings suggest that ${\cal B}(\Omega_b\to J/\psi\Omega(2^2S_{1/2^+},2^4S_{3/2^+}))$ are as large as $(1.2,5.1)\times 10^{-4}$, making them accessible to experiments at LHCb., Comment: 7 pages, 3 figures

Since the emergence of Large Language Models (LLMs), the challenge of effectively leveraging their potential in healthcare has taken center stage. A critical barrier to using LLMs for extracting insights from unstructured clinical notes lies in the prompt engineering process. Despite its pivotal role in determining task performance, a clear framework for prompt optimization remains absent. Current methods to address this gap take either a manual prompt refinement approach, where domain experts collaborate with prompt engineers to create an optimal prompt, which is time-intensive and difficult to scale, or through employing automatic prompt optimizing approaches, where the value of the input of domain experts is not fully realized. To address this, we propose StructEase, a novel framework that bridges the gap between automation and the input of human expertise in prompt engineering. A core innovation of the framework is SamplEase, an iterative sampling algorithm that identifies high-value cases where expert feedback drives significant performance improvements. This approach minimizes expert intervention, to effectively enhance classification outcomes. This targeted approach reduces labeling redundancy, mitigates human error, and enhances classification outcomes. We evaluated the performance of StructEase using a dataset of de-identified clinical narratives from the US National Electronic Injury Surveillance System (NEISS), demonstrating significant gains in classification performance compared to current methods. Our findings underscore the value of expert integration in LLM workflows, achieving notable improvements in F1 score while maintaining minimal expert effort. By combining transparency, flexibility, and scalability, StructEase sets the foundation for a framework to integrate expert input into LLM workflows in healthcare and beyond.

We report the results of an extensive numerical study of the $Sp(4)$ lattice gauge theory with three (Dirac) flavors of fermion in the two-index antisymmetric representation. In the presence of (degenerate) fermion masses, the theory has an enhanced global $SU(6)$ symmetry, broken explicitly and spontaneously to its $SO(6)$ subgroup. This symmetry breaking pattern makes the theory interesting for applications in the context of composite Higgs models, as well as for the implementation of top partial compositeness. It can also provide a dynamical realisation of the strongly interacting massive particle paradigm for the origin of dark matter. We adopt the standard plaquette gauge action with the Wilson-Dirac formulation for the fermions and apply the (rational) hybrid Monte Carlo algorithm in our ensemble generation process. We monitor the autocorrelation and topology of the ensembles. We explore the bare parameter space, and identify the weak and strong coupling regimes separated by a line of first-order bulk phase transitions. We measure two-point correlation functions between meson operators that transform as non-trivial representations of $SO(6)$, and extract the ground-state masses and the decay constants, in all accessible spin and parity channels. In addition, we measure the mass of the first excited state for the vector meson by solving a generalised eigenvalue problem. Spectral quantities show a mass dependence that is compatible with the expectation that, at long distances, the theory undergoes confinement, accompanied by the spontaneous breaking of the approximate global symmetries acting on the matter fields. Finally, we discuss the continuum and massless extrapolations, after setting the physical scale using the gradient flow method, and compare the results to those of existing studies in the quenched approximation, as well as to the literature on closely related theories., Comment: 43 pages, 20 figures, 11 tables

In this study, we aim to determine and solve the deficiency of Stable Diffusion Inpainting (SDI) in following the instruction of both prompt and mask. Due to the training bias from masking, the inpainting quality is hindered when the prompt instruction and image condition are not related. Therefore, we conduct a detailed analysis of the internal representations learned by SDI, focusing on how the mask input influences the cross-attention layer. We observe that adapting text key tokens toward the input mask enables the model to selectively paint within the given area. Leveraging these insights, we propose FreeCond, which adjusts only the input mask condition and image condition. By increasing the latent mask value and modifying the frequency of image condition, we align the cross-attention features with the model's training bias to improve generation quality without additional computation, particularly when user inputs are complicated and deviate from the training setup. Extensive experiments demonstrate that FreeCond can enhance any SDI-based model, e.g., yielding up to a 60% and 58% improvement of SDI and SDXLI in the CLIP score.

Linear block transform coding remains a fundamental component of image and video compression. Although the Discrete Cosine Transform (DCT) is widely employed in all current compression standards, its sub-optimality has sparked ongoing research into discovering more efficient alternative transforms even for fields where it represents a consolidated tool. In this paper, we introduce a novel linear block transform called the Rate Distortion Learned Transform (RDLT), a data-driven transform specifically designed to minimize the rate-distortion (RD) cost when approximating residual blocks. Our approach builds on the latest end-to-end learned compression frameworks, adopting back-propagation and stochastic gradient descent for optimization. However, unlike the nonlinear transforms used in variational autoencoder (VAE)-based methods, the goal is to create a simpler yet optimal linear block transform, ensuring practical integration into existing image and video compression standards. Differently from existing data-driven methods that design transforms based on sample covariance matrices, such as the Karhunen-Lo\`eve Transform (KLT), the proposed RDLT is directly optimized from an RD perspective. Experimental results show that this transform significantly outperforms the DCT or other existing data-driven transforms. Additionally, it is shown that when simulating the integration of our RDLT into a VVC-like image compression framework, the proposed transform brings substantial improvements. All the code used in our experiments has been made publicly available at [1]., Comment: An abstract version of this paper has been accepted at the 2025 Data Compression Conference (DCC)

We report progress on our lattice calculations for the mass spectra of low-lying composite states in the Sp(4) gauge theory coupled to two and three flavors of Dirac fermions transforming in the fundamental and the two-index antisymmetric representations, respectively. This theory provides an ultraviolet completion to the composite Higgs model with Goldstone modes in the SU(4)/Sp(4) coset and with partial compositeness for generating the top-quark mass. We measure the meson and chimera baryon masses. These masses are crucial for constructing the composite Higgs model. In particular, the chimera baryon masses are important inputs for implementing top partial compositeness. We employ Wilson fermions and the Wilson plaquette action in our simulations. Techniques such as APE and Wuppertal smearing, as well as the procedure of generalised eigenvalue problem, are implemented in our analysis., Comment: 12 pages, 8 figures, 1 tables, Proceedings of the 41st International Symposium on Lattice Field Theory (Lattice 2024), July 28th - August 3rd, 2024, University of Liverpool, UK

A broadcasting problem in heterogeneous tree networks with edge weight uncertainty under the postal model is considered in this paper. The broadcasting problem asks for a minmax-regret broadcast center, which minimizes the worst-case loss in the objective function. Due to the presence of edge weight uncertainty, it is not easy to attack the broadcasting problem. An $O(n \log n \log \log n)$-time algorithm is proposed for solving the broadcasting problem.

Objectives: This study examined whether currently existing clinical structural magnetic resonance imaging (MRI) and fluorodeoxyglucose positron emission tomography (18FDG-PET) capabilities and board-certified radiologists' reports and interpretations can assist with traumatic encephalopathy syndrome (TES) diagnosis. Design: retrospective case series. Setting: this study assessed six patients with TES criteria recruited from the Sydney area, Australia Main outcomes: patients' clinical history and clinical presentation along TES diagnostic criteria, board-certified radiologist reports of structural MRI and 18FDG-PET. Results: one patient was classified as possible CTE, and the others were classified as probable CTE with significant RHI exposure history and a spectrum of cognitive deficits and other neuropsychiatric disturbances consistent with TES diagnostic criteria. Most common radiological features included atrophy of posterior superior parietal region and Evans Index > 0.25. FDG-PET's pattern of common regions of hypometabolism and differing hypometabolism across participants suggested that FDG-PET and structural MRI have the potential for stratifying different TES stages. Conclusions: this study highlights the potential of a combined clinical and radiological approach using solely current capabilities to improve TES diagnosis and suggests a larger study. Such expanded investigation is crucial for advancing the ante-mortem diagnosis and management of CTE., Comment: 19 pages, 2 figures, 5 tables

Foundation models refer to artificial intelligence (AI) models that are trained on massive amounts of data and demonstrate broad generalizability across various tasks with high accuracy. These models offer versatile, one-for-many or one-for-all solutions, eliminating the need for developing task-specific AI models. Examples of such foundation models include the Chat Generative Pre-trained Transformer (ChatGPT) and the Segment Anything Model (SAM). These models have been trained on millions to billions of samples and have shown wide-ranging and accurate applications in numerous tasks such as text processing (using ChatGPT) and natural image segmentation (using SAM). In medical image segmentation - finding target regions in medical images - there is a growing need for these one-for-many or one-for-all foundation models. Such models could obviate the need to develop thousands of task-specific AI models, which is currently standard practice in the field. They can also be adapted to tasks with datasets too small for effective training. We discuss two paths to achieve foundation models for medical image segmentation and comment on progress, challenges, and opportunities. One path is to adapt or fine-tune existing models, originally developed for natural images, for use with medical images. The second path entails building models from scratch, exclusively training on medical images.

Neural contextual biasing allows speech recognition models to leverage contextually relevant information, leading to improved transcription accuracy. However, the biasing mechanism is typically based on a cross-attention module between the audio and a catalogue of biasing entries, which means computational complexity can pose severe practical limitations on the size of the biasing catalogue and consequently on accuracy improvements. This work proposes an approximation to cross-attention scoring based on vector quantization and enables compute- and memory-efficient use of large biasing catalogues. We propose to use this technique jointly with a retrieval based contextual biasing approach. First, we use an efficient quantized retrieval module to shortlist biasing entries by grounding them on audio. Then we use retrieved entries for biasing. Since the proposed approach is agnostic to the biasing method, we investigate using full cross-attention, LLM prompting, and a combination of the two. We show that retrieval based shortlisting allows the system to efficiently leverage biasing catalogues of several thousands of entries, resulting in up to 71% relative error rate reduction in personal entity recognition. At the same time, the proposed approximation algorithm reduces compute time by 20% and memory usage by 85-95%, for lists of up to one million entries, when compared to standard dot-product cross-attention., Comment: 13 pages, 7 figures, submitted to IEEE/ACM Transactions on Audio, Speech, and Language Processing

Consider a compact torsion free CR manifold $X$ and assume that $X$ admits a compact CR Lie group action $G$. Let $L$ be a $G$-equivariant rigid CR line bundle over $X$. It seems natural to consider the space of $G$-invariant CR sections in the high tensor powers as quantization space, on which a certain weighted $G$-invariant Fourier-Szeg\H{o} operator projects. Under certain natural assumptions, we show that the group invariant Fourier-Szeg\H{o} projector admits a full asymptotic expansion. As an application, if the tensor power of the line bundle is large enough, we prove that quantization commutes with reduction., Comment: Typos corrected, details added

Most learned B-frame codecs with hierarchical temporal prediction suffer from the domain shift issue caused by the discrepancy in the Group-of-Pictures (GOP) size used for training and test. As such, the motion estimation network may fail to predict large motion properly. One effective strategy to mitigate this domain shift issue is to downsample video frames for motion estimation. However, finding the optimal downsampling factor involves a time-consuming rate-distortion optimization process. This work introduces lightweight classifiers to determine the downsampling factor. To strike a good rate-distortion-complexity trade-off, our classifiers observe simple state signals, including only the coding and reference frames, to predict the best downsampling factor. We present two variants that adopt binary and multi-class classifiers, respectively. The binary classifier adopts the Focal Loss for training, classifying between motion estimation at high and low resolutions. Our multi-class classifier is trained with novel soft labels incorporating the knowledge of the rate-distortion costs of different downsampling factors. Both variants operate as add-on modules without the need to re-train the B-frame codec. Experimental results confirm that they achieve comparable coding performance to the brute-force search methods while greatly reducing computational complexity.

In this paper, we first show that current learning-based video codecs, specifically the SSF codec, are not suitable for real-world applications due to the mismatch between the encoder and decoder caused by floating-point round-off errors. To address this issue, we propose the static quantization of the hyper prior decoding path. The quantization parameters are determined through an exhaustive search of all possible combinations of observers and quantization schemes from PyTorch. For the SSF codec, when encoding and decoding on different machines, the proposed solution effectively mitigates the mismatch issue and enhances compression efficiency results by preventing severe image quality degradation. When encoding and decoding are performed on the same machine, it constrains the average BD-rate increase to 9.93% and 9.02% for UVG and HEVC-B sequences, respectively., Comment: 4 pages, submitted to ISCAS2025

The proliferation of machine learning models in diverse clinical applications has led to a growing need for high-fidelity, medical image training data. Such data is often scarce due to cost constraints and privacy concerns. Alleviating this burden, medical image synthesis via generative adversarial networks (GANs) emerged as a powerful method for synthetically generating photo-realistic images based on existing sets of real medical images. However, the exact image set size required to efficiently train such a GAN is unclear. In this work, we experimentally establish benchmarks that measure the relationship between a sample dataset size and the fidelity of the generated images, given the dataset's distribution of image complexities. We analyze statistical metrics based on delentropy, an image complexity measure rooted in Shannon's entropy in information theory. For our pipeline, we conduct experiments with two state-of-the-art GANs, StyleGAN 3 and SPADE-GAN, trained on multiple medical imaging datasets with variable sample sizes. Across both GANs, general performance improved with increasing training set size but suffered with increasing complexity., Comment: Accepted to ACCV, Workshop on Generative AI for Synthetic Medical Data

In this Letter, we investigate the turbulence and energy injection in the extended nebulae surrounding two luminous obscured quasars, WISEA J100211.29$+$013706.7 ($z=1.5933$) and SDSS J165202.64$+$172852.3 ($z=2.9489$). Utilizing high-resolution data from the NIRSpec IFU onboard the James Webb Space Telescope, we analyze the velocity fields of line-emitting gas in and around these quasars and construct the second-order velocity structure functions (VSFs) to quantify turbulent motions across different spatial scales. Our findings reveal a notable flattening in the VSFs from $\approx\!3$ kpc up to a scale of 10--20 kpc, suggesting that energy injection predominantly occurs at a scale $\lesssim$10 kpc, likely powered by quasar outflows and jet-driven bubbles. The extended spatial range of flat VSFs may also indicate the presence of multiple energy injection sources at these scales. For J1652, the turbulent energy in the host interstellar medium (ISM) is significantly higher than in tidally stripped gas, consistent with the expectation of active galactic nucleus (AGN) activities stirring up the host ISM. Compared to the VSFs observed on spatial scales of 10--50 kpc around lower-redshift UV-bright quasars, these obscured quasars exhibit higher turbulent energies in their immediate surroundings, implying different turbulence drivers between the ISM and halo-scale gas. Future studies with an expanded sample are essential to elucidate further the extent and the pivotal role of AGNs in shaping the gas kinematics of host galaxies and beyond., Comment: 10 pages, 4 figures; submitted; comments welcome!

Quasar feedback may play a key role in the evolution of massive galaxies. The dust-reddened quasar, F2M110648.35$+$480712 at $z = 0.4352$ is one of the few cases at its redshift that exhibits powerful quasar feedback through bipolar outflows. Our new observation with the integral field unit mode of Near-infrared Spectrograph onboard JWST opens a new window to examine this spectacular outflow through Pa$\alpha$ emission line with $\sim$3$\times$ better spatial resolution than previous work. The morphology and kinematics of the Pa$\alpha$ nebula confirm the existence of a bipolar outflow extending on a scale of $\sim$17$\times$14 kpc and with a velocity reaching $\sim$1100 km s$^{-1}$. The higher spatial resolution of our new observation leads to more reliable measurements of outflow kinematics. Considering only the spatially resolved outflow and assuming an electron density of 100 cm$^{-2}$, the mass, momentum and kinetic energy outflow rates are $\sim$50-210 M$_{\odot}$ yr$^{-1}$, $\sim$0.3-1.7$\times$10$^{36}$ dynes ($\sim$14-78\% of the quasar photon momentum flux) and $\sim$0.16-1.27$\times$10$^{44}$ erg s$^{-1}$ ($\sim$0.02-0.20\% of the quasar bolometric luminosity), respectively. The local instantaneous outflow rates generally decrease radially. We infer that the quasar is powerful enough to drive the outflow, while stellar processes cannot be overlooked as a contributing energy source. The mass outflow rate is $\sim$0.4-1.5 times the star formation rate, and the ratio of kinetic energy outflow rate to the quasar bolometric luminosity is comparable to the minimum value required for negative quasar feedback in simulations. This outflow may help regulate the star formation activity within the system to some extent., Comment: 14 pages, 6 figures, 1 table, ApJ in review

We measure the masses of the pseudoscalar flavour-singlet meson states in the $Sp(4)$ gauge theory coupled to two Dirac fermions transforming in the fundamental representation and three Dirac fermions in the antisymmetric representation. This theory provides a compelling ultraviolet completion for the minimal composite Higgs model implementing also partial compositeness for the top quark. The spectrum contains two, comparatively light, pseudoscalar flavour-singlet states, which mix with one another. One of them is a Nambu-Goldstone boson (in the massless limit), whereas the other receives a mass from the $U(1)_A$ axial anomaly. We demonstrate how to measure the mixing between these two states. For moderately heavy fermion masses, we find that the two wave functions are dominated by one of the fermion representations, mixing effects being small., Comment: 12 pages, 2 figures, 2 tables, Proceedings of the 41st International Symposium on Lattice Field Theory (Lattice 2024), July 28th - August 3rd, 2024, University of Liverpool, UK

Spectral densities encode non-perturbative information crucial in computing physical observables in strongly coupled field theories. Using lattice gauge theory data, we perform a systematic study to demonstrate the potential of recent technological advances in the reconstruction of spectral densities. We develop, maintain and make publicly available dedicated analysis code that can be used for broad classes of lattice theories. As a test case, we analyse the Sp(4) gauge theory coupled to an admixture of fermions transforming in the fundamental and two-index antisymmetric representations. We measure the masses of mesons in energy-smeared spectral densities, after optimising the smearing parameters for available lattice ensembles. We present a summary of the mesons mass spectrum in all the twelve (flavored) channels available, including also several excited states., Comment: 9 pages, 5 figures, contribution for the 41th International Symposium on Lattice Field Theory (Lattice 2024), 28 July - 3 August 2024, Liverpool, UK, presenting the results of the paper: arXiv:2405.01388

This paper presents an innovative large language model (LLM) agent framework for enhancing diagnostic accuracy in simulated clinical environments using the AgentClinic benchmark. The proposed automatic correction enables doctor agents to iteratively refine their reasoning and actions following incorrect diagnoses, fostering improved decision-making over time. Experiments show that the implementation of the adaptive LLM-based doctor agents achieve correct diagnoses through dynamic interactions with simulated patients. The evaluations highlight the capacity of autonomous agents to adapt and improve in complex medical scenarios. Future enhancements will focus on refining the algorithm and expanding its applicability across a wider range of tasks and different large language models.

Current Artificial Intelligence (AI) computation systems face challenges, primarily from the memory-wall issue, limiting overall system-level performance, especially for Edge devices with constrained battery budgets, such as smartphones, wearables, and Internet-of-Things sensor systems. In this paper, we propose a new SRAM-based Compute-In-Memory (CIM) accelerator optimized for Spiking Neural Networks (SNNs) Inference. Our proposed architecture employs a multiport SRAM design with multiple decoupled Read ports to enhance the throughput and Transposable Read-Write ports to facilitate online learning. Furthermore, we develop an Arbiter circuit for efficient data-processing and port allocations during the computation. Results for a 128$\times$128 array in 3nm FinFET technology demonstrate a 3.1$\times$ improvement in speed and a 2.2$\times$ enhancement in energy efficiency with our proposed multiport SRAM design compared to the traditional single-port design. At system-level, a throughput of 44 MInf/s at 607 pJ/Inf and 29mW is achieved., Comment: DAC 2024 Research Manuscript

Type II supernovae (SNeII) mark the endpoint in the lives of hydrogen-rich massive stars. Their large explosion energies and luminosities allow us to measure distances, metallicities, and star formation rates into the distant Universe. To fully exploit their use in answering different astrophysical problems, high-quality low-redshift data sets are required. Such samples are vital to understand the physics of SNeII, but also to serve as calibrators for distinct - and often lower-quality - samples. We present uBgVri optical and YJH near-infrared (NIR) photometry for 94 low-redshift SNeII observed by the Carnegie Supernova Project (CSP). A total of 9817 optical and 1872 NIR photometric data points are released, leading to a sample of high-quality SNII light curves during the first ~150 days post explosion on a well-calibrated photometric system. The sample is presented and its properties are analysed and discussed through comparison to literature events. We also focus on individual SNeII as examples of classically defined subtypes and outlier objects. Making a cut in the plateau decline rate of our sample (s2), a new subsample of fast-declining SNeII is presented. The sample has a median redshift of 0.015, with the nearest event at 0.001 and the most distant at 0.07. At optical wavelengths (V), the sample has a median cadence of 4.7 days over the course of a median coverage of 80 days. In the NIR (J), the median cadence is 7.2 days over the course of 59 days. The fast-declining subsample is more luminous than the full sample and shows shorter plateau phases. Of the non-standard SNeII highlighted, SN2009A particularly stands out with a steeply declining then rising light curve, together with what appears to be two superimposed P-Cygni profiles of H-alpha in its spectra. We outline the significant utility of these data, and finally provide an outlook of future SNII science., Comment: Accepted for publication in A&A. Photometric data will be uploaded to the CDS and the CSP website, and can also be requested from the first author

New privacy concerns arise with chatbots on group messaging platforms. Chatbots may access information beyond their intended functionalities, such as messages unintended for chatbots or sender's identities. Chatbot operators may exploit such information to infer personal information and link users across groups, potentially leading to personal data breaches, pervasive tracking, and targeted advertising. Our analysis of conversation datasets shows that (1) chatbots often access far more messages than needed, and (2) when a user joins a new group with chatbots, there is a 3.4% chance that at least one of the chatbots can recognize and associate the user with their previous interactions in other groups. Although state-of-the-art group messaging protocols provide robust end-to-end security and some platforms have implemented policies to limit chatbot access, no platforms successfully combine these features. This paper introduces SnoopGuard, a secure group messaging protocol that ensures user privacy against chatbots while maintaining strong end-to-end security. Our method offers selective message access, preventing chatbots from accessing unrelated messages, and ensures sender anonymity within the group. SnoopGuard achieves $O(\log n + m)$ message-sending complexity for a group of $n$ users and $m$ chatbots, compared to $O(\log(n + m))$ in state-of-the-art protocols, with acceptable overhead for enhanced privacy. Our prototype implementation shows that sending a message in a group of 50 users and 10 chatbots takes about 30 milliseconds when integrated with Message Layer Security (MLS)., Comment: 18 pages, 5 figures

In this paper, we propose an efficient transformer architecture that uses reinforced positional embedding to obtain superior performance with half the number of encoder decoder layers. We demonstrate that concatenating positional encoding with trainable token embeddings, normalizing columns in the token embedding matrix, and using the normalized token embedding matrix as the value of the attention layer improve the training and validation loss and the training time in an encoder-decoder Transformer model for a Portuguese-English translation task with 10 epochs or 12 hours of training across 10 trials. Our method, with roughly a threefold parameter reduction compared to the baseline model, yields a mean training loss of 1.21, a mean validation loss of 1.51, and an average training time of 1352.27 seconds per epoch, surpassing the baseline model with the same embedding dimension that employs addition of positional encoding and token embeddings, which achieves a mean training loss of 1.96, a validation loss of 2.18, and an average training time of 4297.79 seconds per epoch. Additionally, we evaluated our proposed architecture and the baseline across 14 diverse translation datasets from TensorFlow. The results indicate that our method consistently achieves lower or comparable training and validation losses, suggesting enhanced learning efficiency.

This paper aims to delve into the rate-distortion-complexity trade-offs of modern neural video coding. Recent years have witnessed much research effort being focused on exploring the full potential of neural video coding. Conditional autoencoders have emerged as the mainstream approach to efficient neural video coding. The central theme of conditional autoencoders is to leverage both spatial and temporal information for better conditional coding. However, a recent study indicates that conditional coding may suffer from information bottlenecks, potentially performing worse than traditional residual coding. To address this issue, recent conditional coding methods incorporate a large number of high-resolution features as the condition signal, leading to a considerable increase in the number of multiply-accumulate operations, memory footprint, and model size. Taking DCVC as the common code base, we investigate how the newly proposed conditional residual coding, an emerging new school of thought, and its variants may strike a better balance among rate, distortion, and complexity., Comment: Accepted to MMSP 2024