Search

Assigning upper elementary grade teachers to teach their strongest subjects to multiple classes ("departmentalizing"), rather than teaching all subjects to a single class, ("self-contained instruction") could mean more specialized instructional expertise in the classroom or focus for teacher planning time and professional development. This study examined the experiences of 90 schools that either voluntarily switched to departmentalized instruction for up to two years or continued with self-contained teaching in 4th and 5th grade classes beginning in fall 2019. The findings show that schools that switched were able to implement the key steps needed to departmentalize instruction but struggled with aspects of the approach. Teachers' reported experiences were consistent with some of both the benefits and challenges that prior research had hypothesized. Although the pandemic significantly disrupted instruction and the study activities, school and teacher experiences during this challenging time may inform schools and districts considering adopting departmentalized instruction.

This appendix is a companion to the report "Evaluation of Departmentalized Instruction in Elementary Schools: Exploring Implementation Experiences." The appendix provides additional information about the study, which sought to explore schools' implementation of departmentalized instruction, teachers' experiences with the strategy, and how the strategy was related to key outcomes. This additional information includes details on departmentalized instruction (Section A), study design and execution (Section B), key findings presented in the report (Section C), and additional findings that are not included in the report but provide important context (Section D).

State Space Models (SSMs) have emerged as an appealing alternative to Transformers for large language models, achieving state-of-the-art accuracy with constant memory complexity which allows for holding longer context lengths than attention-based networks. The superior computational efficiency of SSMs in long sequence modeling positions them favorably over Transformers in many scenarios. However, improving the efficiency of SSMs on request-intensive cloud-serving and resource-limited edge applications is still a formidable task. SSM quantization is a possible solution to this problem, making SSMs more suitable for wide deployment, while still maintaining their accuracy. Quantization is a common technique to reduce the model size and to utilize the low bit-width acceleration features on modern computing units, yet existing quantization techniques are poorly suited for SSMs. Most notably, SSMs have highly sensitive feature maps within the selective scan mechanism (i.e., linear recurrence) and massive outliers in the output activations which are not present in the output of token-mixing in the self-attention modules. To address this issue, we propose a static 8-bit per-tensor SSM quantization method which suppresses the maximum values of the input activations to the selective SSM for finer quantization precision and quantizes the output activations in an outlier-free space with Hadamard transform. Our 8-bit weight-activation quantized Mamba 2.8B SSM benefits from hardware acceleration and achieves a 1.72x lower generation latency on an Nvidia Orin Nano 8G, with only a 0.9% drop in average accuracy on zero-shot tasks. The experiments demonstrate the effectiveness and practical applicability of our approach for deploying SSM-based models of all sizes on both cloud and edge platforms.

Research into community content moderation often assumes that moderation teams govern with a single, unified voice. However, recent work has found that moderators disagree with one another at modest, but concerning rates. The problem is not the root disagreements themselves. Subjectivity in moderation is unavoidable, and there are clear benefits to including diverse perspectives within a moderation team. Instead, the crux of the issue is that, due to resource constraints, moderation decisions end up being made by individual decision-makers. The result is decision-making that is inconsistent, which is frustrating for community members. To address this, we develop Venire, an ML-backed system for panel review on Reddit. Venire uses a machine learning model trained on log data to identify the cases where moderators are most likely to disagree. Venire fast-tracks these cases for multi-person review. Ideally, Venire allows moderators to surface and resolve disagreements that would have otherwise gone unnoticed. We conduct three studies through which we design and evaluate Venire: a set of formative interviews with moderators, technical evaluations on two datasets, and a think-aloud study in which moderators used Venire to make decisions on real moderation cases. Quantitatively, we demonstrate that Venire is able to improve decision consistency and surface latent disagreements. Qualitatively, we find that Venire helps moderators resolve difficult moderation cases more confidently. Venire represents a novel paradigm for human-AI content moderation, and shifts the conversation from replacing human decision-making to supporting it.

In this work, we re-formulate the model compression problem into the customized compensation problem: Given a compressed model, we aim to introduce residual low-rank paths to compensate for compression errors under customized requirements from users (e.g., tasks, compression ratios), resulting in greater flexibility in adjusting overall capacity without being constrained by specific compression formats. However, naively applying SVD to derive residual paths causes suboptimal utilization of the low-rank representation capacity. Instead, we propose Training-free Eigenspace Low-Rank Approximation (EoRA), a method that directly minimizes compression-induced errors without requiring gradient-based training, achieving fast optimization in minutes using a small amount of calibration data. EoRA projects compression errors into the eigenspace of input activations, leveraging eigenvalues to effectively prioritize the reconstruction of high-importance error components. Moreover, EoRA can be seamlessly integrated with fine-tuning and quantization to further improve effectiveness and efficiency. EoRA consistently outperforms previous methods in compensating errors for compressed LLaMA2/3 models on various tasks, such as language generation, commonsense reasoning, and math reasoning tasks (e.g., 31.31%/12.88% and 9.69% improvements on ARC-Easy/ARC-Challenge and MathQA when compensating LLaMA3-8B that is quantized to 4-bit and pruned to 2:4 sparsity). EoRA offers a scalable, training-free solution to compensate for compression errors, making it a powerful tool to deploy LLMs in various capacity and efficiency requirements.

Trajectory anomaly detection is crucial for effective decision-making in urban and human mobility management. Existing methods of trajectory anomaly detection generally focus on training a trajectory generative model and evaluating the likelihood of reconstructing a given trajectory. However, previous work often lacks important contextual information on the trajectory, such as the agent's information (e.g., agent ID) or geographic information (e.g., Points of Interest (POI)), which could provide additional information on accurately capturing anomalous behaviors. To fill this gap, we propose a context-aware anomaly detection approach that models contextual information related to trajectories. The proposed method is based on a trajectory reconstruction framework guided by contextual factors such as agent ID and contextual POI embedding. The injection of contextual information aims to improve the performance of anomaly detection. We conducted experiments in two cities and demonstrated that the proposed approach significantly outperformed existing methods by effectively modeling contextual information. Overall, this paper paves a new direction for advancing trajectory anomaly detection.

Historical maps are invaluable sources of information about the past, and scanned historical maps are increasingly accessible in online libraries. To retrieve maps from these large libraries that contain specific places of interest, previous work has applied computer vision techniques to recognize words on historical maps, enabling searches for maps that contain specific place names. However, searching for multiword place names is challenging due to complex layouts of text labels on historical maps. This paper proposes an efficient query method for searching a given multiword place name on historical maps. Using existing methods to recognize words on historical maps, we link single-word text labels into potential multiword phrases by constructing minimum spanning trees. These trees aim to link pairs of text labels that are spatially close and have similar height, angle, and capitalization. We then query these trees for the given multiword place name. We evaluate the proposed method in two experiments: 1) to evaluate the accuracy of the minimum spanning tree approach at linking multiword place names and 2) to evaluate the number and time range of maps retrieved by the query approach. The resulting maps reveal how places using multiword names have changed on a large number of maps from across history., Comment: 4 pages, 4 figures, and 2 tables. To be published in proceedings ACM SIGSPATIAL 2024 GeoSearch Workshop

We introduce a new benchmark for reward models that quantifies their ability to produce strong language models through RLHF (Reinforcement Learning from Human Feedback). The gold-standard approach is to run a full RLHF training pipeline and directly probe downstream LLM performance. However, this process is prohibitively expensive. To address this, we build a predictive model of downstream LLM performance by evaluating the reward model on proxy tasks. These proxy tasks consist of a large-scale human preference and a verifiable correctness preference dataset, in which we measure 12 metrics across 12 domains. To investigate which reward model metrics are most correlated to gold-standard RLHF outcomes, we launch an end-to-end RLHF experiment on a large-scale crowdsourced human preference platform to view real reward model downstream performance as ground truth. Ultimately, we compile our data and findings into Preference Proxy Evaluations (PPE), the first reward model benchmark explicitly linked to post-RLHF real-world human preference performance, which we open-source for public use and further development. Our code and evaluations can be found at https://github.com/lmarena/PPE .

The ability to locate an object in an image according to natural language instructions is crucial for many real-world applications. In this work we propose LocateBench, a high-quality benchmark dedicated to evaluating this ability. We experiment with multiple prompting approaches, and measure the accuracy of several large vision language models. We find that even the accuracy of the strongest model, GPT-4o, lags behind human accuracy by more than 10%., Comment: We release the dataset at https://usc-tamagotchi.github.io/locate-bench/

Walking as a form of active travel is essential in promoting sustainable transport. It is thus crucial to accurately predict pedestrian crossing intention and avoid collisions, especially with the advent of autonomous and advanced driver-assisted vehicles. Current research leverages computer vision and machine learning advances to predict near-misses; however, this often requires high computation power to yield reliable results. In contrast, this work proposes a low-complexity ensemble-learning approach that employs contextual data for predicting the pedestrian's intent for crossing. The pedestrian is first detected, and their image is then compressed using skeleton-ization, and contextual information is added into a stacked ensemble-learning approach. Our experiments on different datasets achieve similar pedestrian intent prediction performance as the state-of-the-art approaches with 99.7% reduction in computational complexity. Our source code and trained models will be released upon paper acceptance

When a hot Jupiter orbits a star whose effective temperature exceeds $\sim$6100 K, its orbit normal tends to be misaligned with the stellar spin axis. Cooler stars have smaller obliquities. The latter may have been damped by hot Jupiters in resonance lock with axisymmetric stellar gravity modes (azimuthal number $m=0$), as has been recently recognized. Here we allow for resonance locks with non-axisymmetric modes, which affect both stellar obliquity and spin frequency. Obliquities damp for all modes ($-2 \leq m \leq 2$). Stars spin up for $m > 0$, and spin down for $m < 0$. We carry out a population synthesis that assumes hot Jupiters form misaligned around both cool and hot stars, and subsequently lock onto modes whose $m$-values yield the highest mode energies for given starting obliquities. Core hydrogen burning enables hot Jupiters to torque low-mass stars, but not high-mass stars, into spin-orbit alignment. Resonance locking plus stellar spin-down from magnetic braking largely reproduces observed obliquities and stellar rotation rates and how they trend with stellar effective temperature and orbital separation., Comment: 12 pages, 8 figures, submitted to AAS Journals

The rise of large language models (LLMs) has significantly impacted various domains, including natural language processing (NLP) and image generation, by making complex computational tasks more accessible. While LLMs demonstrate impressive generative capabilities, there is an ongoing debate about their level of "understanding," particularly in structured domains like SQL. In this paper, we evaluate the extent to which LLMs "understand" SQL by testing them on a series of key SQL tasks. These tasks, such as syntax error detection, missing token identification, query performance prediction, query equivalence checking, and query explanation, assess the models' proficiency in recognition, context awareness, semantics, and coherence, which are essential skills for SQL understanding. We generate labeled datasets from well-known workloads, and evaluate the latest LLMs, focusing on how query complexity and syntactic features influence performance. Our results indicate that while GPT4 excels at tasks requiring recognition and context, all models struggle with deeper semantic understanding and coherence, especially in query equivalence and performance estimation, revealing the limitations of current LLMs in achieving full SQL comprehension., Comment: 12 pages conference submission

The first very long baseline interferometry (VLBI) detections at 870$\mu$m wavelength (345$\,$GHz frequency) are reported, achieving the highest diffraction-limited angular resolution yet obtained from the surface of the Earth, and the highest-frequency example of the VLBI technique to date. These include strong detections for multiple sources observed on inter-continental baselines between telescopes in Chile, Hawaii, and Spain, obtained during observations in October 2018. The longest-baseline detections approach 11$\,$G$\lambda$ corresponding to an angular resolution, or fringe spacing, of 19$\mu$as. The Allan deviation of the visibility phase at 870$\mu$m is comparable to that at 1.3$\,$mm on the relevant integration time scales between 2 and 100$\,$s. The detections confirm that the sensitivity and signal chain stability of stations in the Event Horizon Telescope (EHT) array are suitable for VLBI observations at 870$\mu$m. Operation at this short wavelength, combined with anticipated enhancements of the EHT, will lead to a unique high angular resolution instrument for black hole studies, capable of resolving the event horizons of supermassive black holes in both space and time., Comment: Corresponding author: S. Doeleman

We present resolved $3.6-250~\mu$m dust spectral energy distribution (SED) fitting for $\sim 800$ nearby galaxies. We measure the distribution of radiation field intensities heating the dust, the dust mass surface density ($\Sigma_{\rm d}$), and the fraction of dust in the form of polycyclic aromatic hydrocarbons (PAHs; $q_{\rm PAH}$). We find that the average interstellar radiation field ($\overline{U}$) is correlated both with stellar mass surface density ($\Sigma_{\star}$) and star formation rate surface density ($\Sigma_{\rm SFR}$), while more intense radiation fields are only correlated with $\Sigma_{\rm SFR}$. We show that $q_{\rm PAH}$ is a steeply decreasing function of $\Sigma_{\rm SFR}$, likely reflecting PAH destruction in H II regions. Galaxy integrated $q_{\rm PAH}$ is strongly, negatively correlated with specific star formation rate (sSFR) and offset from the star-forming ``main sequence'' ($\Delta$MS), suggesting that both metallicity and star formation intensity play a role in setting the global $q_{\rm PAH}$. We also find a nearly constant M$_{\rm d}$/M$_\star$ ratio for galaxies on the main sequence, with a lower ratio for more quiescent galaxies, likely due to their lower gas fractions. From these results, we construct prescriptions to estimate the radiation field distribution in both integrated and resolved galaxies. We test these prescriptions by comparing our predicted $\overline{U}$ to results of SED fitting for stacked "main sequence" galaxies at $0

Multimodal learning plays a crucial role in enabling machine learning models to fuse and utilize diverse data sources, such as text, images, and audio, to support a variety of downstream tasks. A unified representation across various modalities is particularly important for improving efficiency and performance. Recent binding methods, such as ImageBind (Girdhar et al., 2023), typically use a fixed anchor modality to align multimodal data in the anchor modal embedding space. In this paper, we mathematically analyze the fixed anchor binding methods and uncover notable limitations: (1) over-reliance on the choice of the anchor modality, (2) failure to capture intra-modal information, and (3) failure to account for inter-modal correlation among non-anchored modalities. To address these limitations, we propose CentroBind, a simple yet powerful approach that eliminates the need for a fixed anchor; instead, it employs dynamically adjustable centroid-based anchors generated from all available modalities, resulting in a balanced and rich representation space. We theoretically demonstrate that our method captures three crucial properties of multimodal learning: intra-modal learning, inter-modal learning, and multimodal alignment, while also constructing a robust unified representation across all modalities. Our experiments on both synthetic and real-world datasets demonstrate the superiority of the proposed method, showing that dynamic anchor methods outperform all fixed anchor binding methods as the former captures more nuanced multimodal interactions.

Model predictive control (MPC) has become the de facto standard action space for local planning and learning-based control in many continuous robotic control tasks, including autonomous driving. MPC solves a long-horizon cost optimization as a series of short-horizon optimizations based on a global planner-supplied reference path. The primary challenge in MPC, however, is that the computational budget for re-planning has a hard limit, which frequently inhibits exact optimization. Modern edge networks provide low-latency communication and heterogeneous properties that can be especially beneficial in this situation. We propose a novel framework for edge-assisted MPC (E-MPC) for path planning that exploits the heterogeneity of edge networks in three important ways: 1) varying computational capacity, 2) localized sensor information, and 3) localized observation histories. Theoretical analysis and extensive simulations are undertaken to demonstrate quantitatively the benefits of E-MPC in various scenarios, including maps, channel dynamics, and availability and density of edge nodes. The results confirm that E-MPC has the potential to reduce costs by a greater percentage than standard MPC does.

The Ly$\alpha$ emission has emerged as a powerful tool for probing diffuse gas within the large-scale structure of the universe. In this paper, we investigate cosmic Ly$\alpha$ emission by post-processing cosmological simulations from \texttt{IllustrisTNG} and \texttt{THESAN} project. Specifically, we calculate the Ly$\alpha$ emission from galaxies, circum-galactic medium (CGM) and inter-galactic medium (IGM) across various redshifts. Our results show that IGM alone is significantly under the current observational upper limits. Meanwhile, CGM overshoots the observed galaxy contribution at $z \lesssim 0.5$ indicating that either the escape fraction for the inner CGM is less than unity or the current photoionization equilibrium treatment with an approximate self-shielding prescription is less accurate. The galaxy component also overshoots at low redshift, indicating that the escape fraction has strong evolution caused by an evolving halo mass function and dust growth distribution, that agrees with observationally inferred escape fractions. Furthermore, our findings suggest that the Ly$\alpha$ emission from diffuse gas (CGM+IGM) peaked at $z \sim 4$ and diminishes toward lower redshift. The Ly$\alpha$ emission from diffuse gas mainly originates through the collisional excitation of hot plasma. By comparing models with observation, our predicted Ly$\alpha$ emission from diffuse gas remains $\sim 6$ times fainter than the observed cosmic Ly$\alpha$ emission at $z=1-3$. However, future large telescopes may hold great promise to detect Ly$\alpha$ emission from diffuse gas toward $z>3$.

In challenging environments with significant noise and reverberation, traditional speech enhancement (SE) methods often lead to over-suppressed speech, creating artifacts during listening and harming downstream tasks performance. To overcome these limitations, we propose a novel approach called Restorative SE (RestSE), which combines a lightweight SE module with a generative codec module to progressively enhance and restore speech quality. The SE module initially reduces noise, while the codec module subsequently performs dereverberation and restores speech using generative capabilities. We systematically explore various quantization techniques within the codec module to optimize performance. Additionally, we introduce a weighted loss function and feature fusion that merges the SE output with the original mixture, particularly at segments where the SE output is heavily distorted. Experimental results demonstrate the effectiveness of our proposed method in enhancing speech quality under adverse conditions. Audio demos are available at: https://sophie091524.github.io/RestorativeSE/., Comment: Paper in submission

It is common to reject undesired outputs of Large Language Models (LLMs); however, current methods to do so require an excessive amount of computation, or severely distort the distribution of outputs. We present a method to balance the distortion of the output distribution with computational efficiency, allowing for the generation of long sequences of text with difficult-to-satisfy constraints, with less amplification of low probability outputs compared to existing methods. We show through a series of experiments that the task-specific performance of our method is comparable to methods that do not distort the output distribution, while being much more computationally efficient., Comment: 9 pages main, 22 pages total

Children often suffer wrist trauma in daily life, while they usually need radiologists to analyze and interpret X-ray images before surgical treatment by surgeons. The development of deep learning has enabled neural networks to serve as computer-assisted diagnosis (CAD) tools to help doctors and experts in medical image diagnostics. Since the You Only Look Once Version-8 (YOLOv8) model has obtained the satisfactory success in object detection tasks, it has been applied to various fracture detection. This work introduces four variants of Feature Contexts Excitation-YOLOv8 (FCE-YOLOv8) model, each incorporating a different FCE module (i.e., modules of Squeeze-and-Excitation (SE), Global Context (GC), Gather-Excite (GE), and Gaussian Context Transformer (GCT)) to enhance the model performance. Experimental results on GRAZPEDWRI-DX dataset demonstrate that our proposed YOLOv8+GC-M3 model improves the mAP@50 value from 65.78% to 66.32%, outperforming the state-of-the-art (SOTA) model while reducing inference time. Furthermore, our proposed YOLOv8+SE-M3 model achieves the highest mAP@50 value of 67.07%, exceeding the SOTA performance. The implementation of this work is available at https://github.com/RuiyangJu/FCE-YOLOv8., Comment: arXiv admin note: text overlap with arXiv:2407.03163

Recent end-to-end speech language models (SLMs) have expanded upon the capabilities of large language models (LLMs) by incorporating pre-trained speech models. However, these SLMs often undergo extensive speech instruction-tuning to bridge the gap between speech and text modalities. This requires significant annotation efforts and risks catastrophic forgetting of the original language capabilities. In this work, we present a simple yet effective automatic process for creating speech-text pair data that carefully injects speech paralinguistic understanding abilities into SLMs while preserving the inherent language capabilities of the text-based LLM. Our model demonstrates general capabilities for speech-related tasks without the need for speech instruction-tuning data, achieving impressive performance on Dynamic-SUPERB and AIR-Bench-Chat benchmarks. Furthermore, our model exhibits the ability to follow complex instructions derived from LLMs, such as specific output formatting and chain-of-thought reasoning. Our approach not only enhances the versatility and effectiveness of SLMs but also reduces reliance on extensive annotated datasets, paving the way for more efficient and capable speech understanding systems., Comment: Submitted to ICASSP 2025

In this study, we examined the impact of recommendation systems' algorithms on individuals' collaborator choices when forming teams. Different algorithmic designs can lead individuals to select one collaborator over another, thereby shaping their teams' composition, dynamics, and performance. To test this hypothesis, we conducted a 2 x 2 between-subject laboratory experiment with 332 participants who assembled teams using a recommendation system. We tested four algorithms that controlled the participants' agency to choose collaborators and the inclusion of fairness criteria. Our results show that participants assigned by an algorithm to work in highly diverse teams struggled to work with different and unfamiliar individuals, while participants enabled by an algorithm to choose collaborators without fairness criteria formed homogenous teams without the necessary skills. In contrast, combining users' agency and fairness criteria in an algorithm enhanced teams' performance and composition. This study breaks new ground by providing insights into how algorithms can augment team formation.

This paper analyzes the impact of causal manner in the text encoder of text-to-image (T2I) diffusion models, which can lead to information bias and loss. Previous works have focused on addressing the issues through the denoising process. However, there is no research discussing how text embedding contributes to T2I models, especially when generating more than one object. In this paper, we share a comprehensive analysis of text embedding: i) how text embedding contributes to the generated images and ii) why information gets lost and biases towards the first-mentioned object. Accordingly, we propose a simple but effective text embedding balance optimization method, which is training-free, with an improvement of 90.05% on information balance in stable diffusion. Furthermore, we propose a new automatic evaluation metric that quantifies information loss more accurately than existing methods, achieving 81% concordance with human assessments. This metric effectively measures the presence and accuracy of objects, addressing the limitations of current distribution scores like CLIP's text-image similarities., Comment: Accepted to NeurIPS 2024

We consider an autonomous vehicle (AV) agent performing a long-term cost-minimization problem in the elapsed time $T$ over sequences of states $s_{1:T}$ and actions $a_{1:T}$ for some fixed, known (though potentially learned) cost function $C(s_t,a_t)$, approximate system dynamics $P$, and distribution over initial states $d_0$. The goal is to minimize the expected cost-to-go of the driving trajectory $\tau = s_1, a_1, ..., s_T, a_T$ from the initial state.

Wrist trauma and even fractures occur frequently in daily life, particularly among children who account for a significant proportion of fracture cases. Before performing surgery, surgeons often request patients to undergo X-ray imaging first, and prepare for the surgery based on the analysis of the X-ray images. With the development of neural networks, You Only Look Once (YOLO) series models have been widely used in fracture detection for Computer-Assisted Diagnosis, where the YOLOv8 model has obtained the satisfactory results. Applying the attention modules to neural networks is one of the effective methods to improve the model performance. This paper proposes YOLOv8-ResCBAM, which incorporates Convolutional Block Attention Module integrated with resblock (ResCBAM) into the original YOLOv8 network architecture. The experimental results on the GRAZPEDWRI-DX dataset demonstrate that the mean Average Precision calculated at Intersection over Union threshold of 0.5 (mAP 50) of the proposed model increased from 63.6% of the original YOLOv8 model to 65.8%, which achieves the state-of-the-art performance. The implementation code is available at https://github.com/RuiyangJu/Fracture_Detection_Improved_YOLOv8., Comment: Accepted by ICONIP 2024. arXiv admin note: substantial text overlap with arXiv:2402.09329

The challenge in reconfigurable manipulation of sound waves using metasurfaces lies in achieving precise control over acoustic behavior while developing efficient and practical tuning methods for structural configurations. However, most studies on reconfigurable acoustic metasurfaces rely on cumbersome and time-consuming control systems. These approaches often struggle with fabrication techniques, as conventional methods face limitations such as restricted material choices, challenges in achieving complex geometries, and difficulties in incorporating flexible components. This paper proposes a novel approach for developing a reconfigurable metasurface inspired by the Kresling origami pattern, designed for programmable manipulation of acoustic waves at an operating frequency of 2000 Hz. The origami unit cell is fabricated using multi-material 3D printing technology, allowing for the simultaneous printing of two materials with different mechanical properties, thus creating a bistable origami-based structure. Through optimization, two equilibrium states achieve a reflection phase difference of {\pi} through the application of small axial force, F, or torque, T. Various configurations of the metasurface, generated from different combinations of these two equilibria, enable distinct reflective behaviors with switchable and programmable functionalities. The principle of this work simplifies the shaping of acoustic waves through a straightforward mechanical mechanism, eliminating the need for complex control systems and time-consuming adjustments. This innovative approach paves a novel and effective perspective for developing on-demand switchable and tunable devices across diverse fields, including electromagnetics, mechanics, and elastics, leveraging multi-material printing technology.

With the prevalence of instant messaging and social media platforms, emojis have become important artifacts for expressing emotions and feelings in our daily lives. We ask how HCI researchers have examined the role and evolution of emojis in sharing emotions over the past 10 years. We conducted a systematic literature review of papers addressing emojis employed for emotion communication between users. After screening more than 1,000 articles, we identified 42 articles of studies analyzing ways and systems that enable users to share emotions with emojis. Two main themes described how these papers have (1) improved how users select the right emoji from an increasing emoji lexicon, and (2) employed emojis in new ways and digital materials to enhance communication. We also discovered an increasingly broad scope of functionality across appearance, medium, and affordance. We discuss and offer insights into potential opportunities and challenges emojis will bring for HCI research.

Timely disease diagnosis is challenging due to increasing disease burdens and limited clinician availability. AI shows promise in diagnosis accuracy but faces real-world application issues due to insufficient validation in clinical workflows and diverse populations. This study addresses gaps in medical AI downstream accountability through a case study on age-related macular degeneration (AMD) diagnosis and severity classification. We designed and implemented an AI-assisted diagnostic workflow for AMD, comparing diagnostic performance with and without AI assistance among 24 clinicians from 12 institutions with real patient data sampled from the Age-Related Eye Disease Study (AREDS). Additionally, we demonstrated continual enhancement of an existing AI model by incorporating approximately 40,000 additional medical images (named AREDS2 dataset). The improved model was then systematically evaluated using both AREDS and AREDS2 test sets, as well as an external test set from Singapore. AI assistance markedly enhanced diagnostic accuracy and classification for 23 out of 24 clinicians, with the average F1-score increasing by 20% from 37.71 (Manual) to 45.52 (Manual + AI) (P-value < 0.0001), achieving an improvement of over 50% in some cases. In terms of efficiency, AI assistance reduced diagnostic times for 17 out of the 19 clinicians tracked, with time savings of up to 40%. Furthermore, a model equipped with continual learning showed robust performance across three independent datasets, recording a 29% increase in accuracy, and elevating the F1-score from 42 to 54 in the Singapore population.

Trajectory forecasting in healthcare data has been an important area of research in precision care and clinical integration for computational methods. In recent years, generative AI models have demonstrated promising results in capturing short and long range dependencies in time series data. While these models have also been applied in healthcare, most of them only predict one value at a time, which is unrealistic in a clinical setting where multiple measures are taken at once. In this work, we extend the framework temporal fusion transformer (TFT), a multi-horizon time series prediction tool, and propose TFT-multi, an end-to-end framework that can predict multiple vital trajectories simultaneously. We apply TFT-multi to forecast 5 vital signs recorded in the intensive care unit: blood pressure, pulse, SpO2, temperature and respiratory rate. We hypothesize that by jointly predicting these measures, which are often correlated with one another, we can make more accurate predictions, especially in variables with large missingness. We validate our model on the public MIMIC dataset and an independent institutional dataset, and demonstrate that this approach outperforms state-of-the-art univariate prediction tools including the original TFT and Prophet, as well as vector regression modeling for multivariate prediction. Furthermore, we perform a study case analysis by applying our pipeline to forecast blood pressure changes in response to actual and hypothetical pressor administration.

Environmental hazards place certain individuals at disproportionately higher risks. As these hazards increasingly endanger human health, precise identification of the most vulnerable population subgroups is critical for public health. Moderated multiple regression (MMR) offers a straightforward method for investigating this by adding interaction terms between the exposure to a hazard and other population characteristics to a linear regression model. However, when the vulnerabilities are hidden within a cross-section of many characteristics, MMR is often limited in its capabilities to find any meaningful discoveries. Here, we introduce a hybrid method, named regression-guided neural networks (ReGNN), which utilizes artificial neural networks (ANNs) to non-linearly combine predictors, generating a latent representation that interacts with a focal predictor (i.e. variable measuring exposure to an environmental hazard). We showcase the use of ReGNN for investigating the population heterogeneity in the health effects of exposure to air pollution (PM2.5) on cognitive functioning scores. We demonstrate that population heterogeneity that would otherwise be hidden using traditional MMR can be found using ReGNN by comparing its results to the fit results of the traditional MMR models. In essence, ReGNN is a novel tool that enhances traditional regression models by effectively summarizing and quantifying an individual's susceptibility to health risks.

Previous work has shown that the languages recognized by average-hard attention transformers (AHATs) and softmax-attention transformers (SMATs) are within the circuit complexity class TC$^0$. However, these results assume limited-precision arithmetic: using floating-point numbers with O(log n) bits (where n is the length of the input string), Strobl showed that AHATs can be approximated in L-uniform TC$^0$, and Merrill and Sabharwal showed that SMATs can be approximated in DLOGTIME-uniform TC$^0$. Here, we improve these results, showing that AHATs with no approximation, SMATs with O(poly(n)) bits of floating-point precision, and SMATs with at most $2^{-O(poly(n))}$ absolute error are all in DLOGTIME-uniform TC$^0$.

Multi-agent strategies have emerged as a promising approach to enhance the reasoning abilities of Large Language Models (LLMs) by assigning specialized roles in the problem-solving process. Concurrently, Tree of Thoughts (ToT) methods have shown potential in improving reasoning for complex question-answering tasks by exploring diverse reasoning paths. A critical limitation in multi-agent reasoning is the 'Reasoner' agent's shallow exploration of reasoning paths. While ToT strategies could help mitigate this problem, they may generate flawed reasoning branches, which could harm the trustworthiness of the final answer. To leverage the strengths of both multi-agent reasoning and ToT strategies, we introduce a novel approach combining ToT-based Reasoner agents with a Thought Validator agent. Multiple Reasoner agents operate in parallel, employing ToT to explore diverse reasoning paths. The Thought Validator then scrutinizes these paths, considering a Reasoner's conclusion only if its reasoning is valid. This method enables a more robust voting strategy by discarding faulty reasoning paths, enhancing the system's ability to tackle tasks requiring systematic and trustworthy reasoning. Our method demonstrates superior performance compared to existing techniques when evaluated on the GSM8K dataset, outperforming the standard ToT strategy by an average 5.6\% across four LLMs.

Edge computing is deemed a promising technique to execute latency-sensitive applications by offloading computation-intensive tasks to edge servers. Extensive research has been conducted in the field of end-device to edge server task offloading for several goals, including latency minimization, energy optimization, and resource optimization. However, few of them consider our mobile computing devices (smartphones, tablets, and laptops) to be edge devices. In this paper, we propose a novel multi-tier edge computing framework, which we refer to as M-TEC, that aims to optimize latency, reduce the probability of failure, and optimize cost while accounting for the sporadic failure of personally owned devices and the changing network conditions. We conduct experiments with a real testbed and a real commercial CBRS 4G network, and the results indicate that M-TEC is capable of reducing the end-to-end latency of applications by at least 8\% compared to the best baseline under a variety of network conditions, while providing reliable performance at an affordable cost., Comment: 12 pages

We explore the potential primordial connection between the black hole merger events detected by LIGO and the nano-Hz stochastic gravitational wave background observed by pulsar timing arrays. We propose an innovative mechanism for the formation of primordial black holes, suggesting that the Poisson fluctuations within the domain wall network can give rise to horizon-sized overdense regions. Our results indicate a plausible common origin for gravitational wave observations in two different frequency bands, potentially linked to the annihilation of the domain wall network at the QCD scale, while accounting for the accretion effects on primordial black holes. Furthermore, we demonstrate that the bias potential induced by the QCD instanton effect may naturally facilitate the annihilation of the domain wall network during the QCD phase transition. Additionally, our scenario can yield the correct axion dark matter relic abundance, particularly if realized within the clockwork axion framework., Comment: 6 pages, 2 figures, 1 table

Recent advances in VLSI fabrication technology have led to die shrinkage and increased layout density, creating an urgent demand for advanced hotspot detection techniques. However, by taking an object detection network as the backbone, recent learning-based hotspot detectors learn to recognize only the problematic layout patterns in the training data. This fact makes these hotspot detectors difficult to generalize to real-world scenarios. We propose a novel lithography simulator-powered hotspot detection framework to overcome this difficulty. Our framework integrates a lithography simulator with an object detection backbone, merging the extracted latent features from both the simulator and the object detector via well-designed cross-attention blocks. Consequently, the proposed framework can be used to detect potential hotspot regions based on I) the variation of possible circuit shape deformation estimated by the lithography simulator, and ii) the problematic layout patterns already known. To this end, we utilize RetinaNet with a feature pyramid network as the object detection backbone and leverage LithoNet as the lithography simulator. Extensive experiments demonstrate that our proposed simulator-guided hotspot detection framework outperforms previous state-of-the-art methods on real-world data., Comment: 14 pages to appear in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems

Domain walls are topological defects produced by the spontaneous symmetry-breaking of discrete symmetry during cosmological phase transitions. The horizon-size domain wall can significantly contribute to the energy density in the late-evolution stage. We propose that the density perturbations from the fluctuations in the number density of the horizon-size domain wall could collapse to form primordial black holes. This mechanism becomes effective when the domain wall energy density ratio to that of the radiation reaches about 0.1 in the radiation-dominated universe. We find that models with $Z_2$ symmetry are excluded for interpreting pulsar timing array observations on the nano-Hz gravitational wave background since this model's domain wall number density fluctuations could lead to an overabundance of the primordial black holes. Moreover, models with $N\sim 10$ domain walls also suffer strong constraints from the overabundance of primordial black holes., Comment: 39 pages, 11 figures, 4 tables

$N{:}M$ sparsity is an emerging model compression method supported by more and more accelerators to speed up sparse matrix multiplication in deep neural networks. Most existing $N{:}M$ sparsity methods compress neural networks with a uniform setting for all layers in a network or heuristically determine the layer-wise configuration by considering the number of parameters in each layer. However, very few methods have been designed for obtaining a layer-wise customized $N{:}M$ sparse configuration for vision transformers (ViTs), which usually consist of transformer blocks involving the same number of parameters. In this work, to address the challenge of selecting suitable sparse configuration for ViTs on $N{:}M$ sparsity-supporting accelerators, we propose ELSA, Exploiting Layer-wise $N{:}M$ Sparsity for ViTs. Considering not only all $N{:}M$ sparsity levels supported by a given accelerator but also the expected throughput improvement, our methodology can reap the benefits of accelerators supporting mixed sparsity by trading off negligible accuracy loss with both memory usage and inference time reduction for ViT models. For instance, our approach achieves a noteworthy 2.9$\times$ reduction in FLOPs for both Swin-B and DeiT-B with only a marginal degradation of accuracy on ImageNet. Our code will be released upon paper acceptance.

We introduce Seed-Music, a suite of music generation systems capable of producing high-quality music with fine-grained style control. Our unified framework leverages both auto-regressive language modeling and diffusion approaches to support two key music creation workflows: controlled music generation and post-production editing. For controlled music generation, our system enables vocal music generation with performance controls from multi-modal inputs, including style descriptions, audio references, musical scores, and voice prompts. For post-production editing, it offers interactive tools for editing lyrics and vocal melodies directly in the generated audio. We encourage readers to listen to demo audio examples at https://team.doubao.com/seed-music "https://team.doubao.com/seed-music"., Comment: Seed-Music technical report, 20 pages, 5 figures

Depth completion aims to predict dense depth maps with sparse depth measurements from a depth sensor. Currently, Convolutional Neural Network (CNN) based models are the most popular methods applied to depth completion tasks. However, despite the excellent high-end performance, they suffer from a limited representation area. To overcome the drawbacks of CNNs, a more effective and powerful method has been presented: the Transformer, which is an adaptive self-attention setting sequence-to-sequence model. While the standard Transformer quadratically increases the computational cost from the key-query dot-product of input resolution which improperly employs depth completion tasks. In this work, we propose a different window-based Transformer architecture for depth completion tasks named Sparse-to-Dense Transformer (SDformer). The network consists of an input module for the depth map and RGB image features extraction and concatenation, a U-shaped encoder-decoder Transformer for extracting deep features, and a refinement module. Specifically, we first concatenate the depth map features with the RGB image features through the input model. Then, instead of calculating self-attention with the whole feature maps, we apply different window sizes to extract the long-range depth dependencies. Finally, we refine the predicted features from the input module and the U-shaped encoder-decoder Transformer module to get the enriching depth features and employ a convolution layer to obtain the dense depth map. In practice, the SDformer obtains state-of-the-art results against the CNN-based depth completion models with lower computing loads and parameters on the NYU Depth V2 and KITTI DC datasets., Comment: Presented at the International Conference on Industrial Automation, Robotics and Control Engineering (IARCE) 2022

While memory-augmented neural networks (MANNs) offer an effective solution for few-shot learning (FSL) by integrating deep neural networks with external memory, the capacity requirements and energy overhead of data movement become enormous due to the large number of support vectors in many-class FSL scenarios. Various in-memory search solutions have emerged to improve the energy efficiency of MANNs. NAND-based multi-bit content addressable memory (MCAM) is a promising option due to its high density and large capacity. Despite its potential, MCAM faces limitations such as a restricted number of word lines, limited quantization levels, and non-ideal effects like varying string currents and bottleneck effects, which lead to significant accuracy drops. To address these issues, we propose several innovative methods. First, the Multi-bit Thermometer Code (MTMC) leverages the extensive capacity of MCAM to enhance vector precision using cumulative encoding rules, thereby mitigating the bottleneck effect. Second, the Asymmetric vector similarity search (AVSS) reduces the precision of the query vector while maintaining that of the support vectors, thereby minimizing the search iterations and improving efficiency in many-class scenarios. Finally, the Hardware-Aware Training (HAT) method optimizes controller training by modeling the hardware characteristics of MCAM, thus enhancing the reliability of the system. Our integrated framework reduces search iterations by up to 32 times, and increases overall accuracy by 1.58% to 6.94%.

In previous work [LM17], Tri Lai and the second author studied a family of subgraphs of the dP3 brane tiling, called Aztec castles, whose dimer partition functions provide combinatorial formulas for cluster variables resulting from mutations of the quiver associated with the del Pezzo surface dP3. In our paper, we investigate a variant of the dP3 quiver by considering a second alphabet of variables that breaks the symmetries of the relevant recurrences. This deformation is motivated by the theory of cluster algebras with principal coefficients introduced by Fomin and Zelevinsky. Our main result gives an explicit formula extending previously known generating functions for dP3 cluster variables by using Aztec castles and constructing their associated minimal matchings., Comment: 26 pages, 53 figures

Photoemission electron microscopy (PEEM) has evolved into an indispensable tool for structural and magnetic characterization of surfaces at the nanometer scale. In strong contrast to synchrotron-radiation-based X-ray PEEM as a leading method for element-specific magnetic properties via magnetic circular dichroism (MCD), laboratory ultraviolet (UV) PEEM has seen limited application with much smaller dichroism effects for in-plane magnetization. Here we introduce darkfield PEEM as a novel approach to enhance MCD contrast in threshold photoemission, enabling efficient MCD imaging with significantly enhanced contrast by an order-of-magnitude for Fe(001). This advancement paves the way for MCD imaging on femtosecond timescales using modern lasers. The experimental results will be quantitatively benchmarked against advanced relativistic photoemission calculations.

Developing a versatile deep neural network to model music audio is crucial in MIR. This task is challenging due to the intricate spectral variations inherent in music signals, which convey melody, harmonics, and timbres of diverse instruments. In this paper, we introduce Mel-RoFormer, a spectrogram-based model featuring two key designs: a novel Mel-band Projection module at the front-end to enhance the model's capability to capture informative features across multiple frequency bands, and interleaved RoPE Transformers to explicitly model the frequency and time dimensions as two separate sequences. We apply Mel-RoFormer to tackle two essential MIR tasks: vocal separation and vocal melody transcription, aimed at isolating singing voices from audio mixtures and transcribing their lead melodies, respectively. Despite their shared focus on singing signals, these tasks possess distinct optimization objectives. Instead of training a unified model, we adopt a two-step approach. Initially, we train a vocal separation model, which subsequently serves as a foundation model for fine-tuning for vocal melody transcription. Through extensive experiments conducted on benchmark datasets, we showcase that our models achieve state-of-the-art performance in both vocal separation and melody transcription tasks, underscoring the efficacy and versatility of Mel-RoFormer in modeling complex music audio signals., Comment: Accepted to appear in ISMIR 2024

Coffee is one of the most valuable primary commodities. Despite this, the common selection technique of green coffee beans relies on personnel visual inspection, which is labor-intensive and subjective. Therefore, an efficient way to evaluate the quality of beans is needed. In this paper, we demonstrate a site-independent approach to find site-specific color features of the seed coat in qualified green coffee beans. We then propose two evaluation schemes for green coffee beans based on this site-specific color feature of qualified beans. Due to the site-specific properties of these color features, machine learning classifiers indicate that compared with the existing evaluation schemes of beans, our evaluation schemes have the advantages of being simple, having less computational costs, and having universal applicability. Finally, this site-specific color feature can distinguish qualified beans from different growing sites. Moreover, this function can prevent cheating in the coffee business and is unique to our evaluation scheme of beans., Comment: 21 pages, 7 figures

We show in this paper that in the context of graded Lie algebras, all cuspidal character sheaves arise from a nearby-cycle construction followed by a Fourier--Sato transform in a very specific manner. Combined with results of the last two named authors, this completes the explicit description of cuspidal character sheaves for Vinberg's type I graded classical Lie algebras.

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

Progress in the task of symbolic music generation may be lagging behind other tasks like audio and text generation, in part because of the scarcity of symbolic training data. In this paper, we leverage the greater scale of audio music data by applying pre-trained MIR models (for transcription, beat tracking, structure analysis, etc.) to extract symbolic events and encode them into token sequences. To the best of our knowledge, this work is the first to demonstrate the feasibility of training symbolic generation models solely from auto-transcribed audio data. Furthermore, to enhance the controllability of the trained model, we introduce SymPAC (Symbolic Music Language Model with Prompting And Constrained Generation), which is distinguished by using (a) prompt bars in encoding and (b) a technique called Constrained Generation via Finite State Machines (FSMs) during inference time. We show the flexibility and controllability of this approach, which may be critical in making music AI useful to creators and users., Comment: ISMIR 2024

Vaporized metal, silicates, and ices on the verge of re-condensing into solid or liquid particles appear in many contexts: behind shocks, in impact ejecta, and within the atmospheres and outflows of stars, disks, planets, and minor bodies. We speculate that a condensing gas might fragment, forming overdensities within relative voids, from a radiation-condensation instability. Seeded with small thermal fluctuations, a condensible gas will exhibit spatial variations in the density of particle condensates. Regions of higher particle density may radiate more, cooling faster. Faster cooling leads to still more condensation, lowering the local pressure. Regions undergoing runaway condensation may collapse under the pressure of their less condensed surroundings. Particle condensates will compactify with collapsing regions, into overdense clumps or macroscopic solids (planetesimals). As a first step toward realizing this hypothetical instability, we calculate the evolution of a small volume of condensing silicate vapor -- a spherical test "bubble" embedded in a background medium whose pressure and radiation field are assumed fixed for simplicity. Such a bubble condenses and collapses upon radiating its latent heat to the background, assuming its energy loss is not stopped by background irradiation. Collapse speeds can range up to sonic, similar to cavitation in terrestrial settings. Adding a non-condensible gas like hydrogen to the bubble stalls the collapse. We discuss whether cavitation can provide a way for mm-sized chondrules and refractory solids to assemble into meteorite parent bodies, focusing on CB/CH chondrites whose constituents likely condensed from silicate/metal vapor released from the most energetic asteroid collisions., Comment: Accepted to ApJ Letters

We propose a simple estimator that allows to calculate the absolute value of a system's partition function from a finite sampling of its canonical ensemble. The estimator utilizes a volume correction term to compensate the effect that the finite sampling cannot cover the whole configuration space. As a proof of concept, the estimator is applied to calculate the partition function for several model systems, and the results are compared with the numerically exact solutions. Excellent agreement is found, demonstrating that a solution for an efficient calculation of partition functions is possible.

We introduce SAM4MLLM, an innovative approach which integrates the Segment Anything Model (SAM) with Multi-Modal Large Language Models (MLLMs) for pixel-aware tasks. Our method enables MLLMs to learn pixel-level location information without requiring excessive modifications to the existing model architecture or adding specialized tokens. We introduce an inquiry-based approach that can effectively find prompt points for SAM to perform segmentation based on MLLM. It combines detailed visual information with the powerful expressive capabilities of large language models in a unified language-based manner without additional computational overhead in learning. Experimental results on pubic benchmarks demonstrate the effectiveness of our approach., Comment: ECCV 2024