Search

The remarkable success of foundation models has sparked growing interest in their application to single-cell biology. Models like Geneformer and scGPT promise to serve as versatile tools in this specialized field. However, representing a cell as a sequence of genes remains an open question since the order of genes is interchangeable. Injecting the gene network graph offers gene relative positions and compact data representation but poses a dilemma: limited receptive fields without in-layer message passing or parameter explosion with message passing in each layer. To pave the way forward, we propose GenoHoption, a new computational framework for single-cell sequencing data that effortlessly combines the strengths of these foundation models with explicit relationships in gene networks. We also introduce a constraint that lightens the model by focusing on learning the predefined graph structure while ensuring further hops are deducted to expand the receptive field. Empirical studies show that our model improves by an average of 1.27% on cell-type annotation and 3.86% on perturbation prediction. Furthermore, our method significantly decreases computational overhead and exhibits few-shot potential. GenoHoption can function as an efficient and expressive bridge, connecting existing single-cell foundation models to gene network graphs.

Lead-halide perovskites are semiconductor materials with attractive properties for photovoltaic and other optoelectronic applications. However, determining crucial electronic material parameters, such as charge-carrier mobility and lifetime, is plagued by a wide range of reported values and inconsistencies caused by interpreting and reporting data originating from different measurement techniques. In this paper, we propose a method for the simultaneous determination of mobility and lifetime using only one technique: transient photoluminescence spectroscopy. By measuring and simulating the decay of the photoluminescence intensity and the redshift of the photoluminescence peak as a function of time after the laser pulse, we extract the mobility, lifetime, and diffusion length of halide perovskite films. With a voltage-dependent steady-state photoluminescence measurement on a cell, we relate the diffusion length to the external voltage and quantify its value at the maximum power point., Comment: 24 pages, 5 figures

Errors are common in time series due to unreliable sensor measurements. Existing methods focus on univariate data but do not utilize the correlation between dimensions. Cleaning each dimension separately may lead to a less accurate result, as some errors can only be identified in the multivariate case. We also point out that the widely used minimum change principle is not always the best choice. Instead, we try to change the smallest number of data to avoid a significant change in the data distribution. In this paper, we propose MTCSC, the constraint-based method for cleaning multivariate time series. We formalize the repair problem, propose a linear-time method to employ online computing, and improve it by exploiting data trends. We also support adaptive speed constraint capturing. We analyze the properties of our proposals and compare them with SOTA methods in terms of effectiveness, efficiency versus error rates, data sizes, and applications such as classification. Experiments on real datasets show that MTCSC can have higher repair accuracy with less time consumption. Interestingly, it can be effective even when there are only weak or no correlations between the dimensions., Comment: 14 pages, 16 figures, conference

Recently, large language models (LLMs) have been successfully applied to many fields, showing outstanding comprehension and reasoning capabilities. Despite their great potential, LLMs usually require dedicated pre-training and fine-tuning for domain-specific applications such as wireless networks. These adaptations can be extremely demanding for computational resources and datasets, while most network devices have limited computation power, and there are a limited number of high-quality networking datasets. To this end, this work explores LLM-enabled wireless networks from the prompt engineering perspective, i.e., designing prompts to guide LLMs to generate desired output without updating LLM parameters. Compared with other LLM-driven methods, prompt engineering can better align with the demands of wireless network devices, e.g., higher deployment flexibility, rapid response time, and lower requirements on computation power. In particular, this work first introduces LLM fundamentals and compares different prompting techniques such as in-context learning, chain-of-thought, and self-refinement. Then we propose two novel prompting schemes for network applications: iterative prompting for network optimization, and self-refined prompting for network prediction. The case studies show that the proposed schemes can achieve comparable performance as conventional machine learning techniques, and our proposed prompting-based methods avoid the complexity of dedicated model training and fine-tuning, which is one of the key bottlenecks of existing machine learning techniques.

The spontaneous conversion of muonium to antimuonium is one of the interesting charged lepton flavor violation phenomena, offering a sensitive probe of potential new physics and serving as a tool to constrain the parameter space beyond the Standard Model. Utilizing a high-intensity muon beam, a Michel electron magnetic spectrometer and a positron transport solenoid together with a positron detection system, MACE aims to discover or constrain this rare process at the conversion probability beyond the level of $10^{-13}$. This report provides an overview of the theoretical framework and detailed experimental design in the search for the muonium-to-antimuonium conversion., Comment: 115 pages, 59 figures

Humanoid robots, with their human-like embodiment, have the potential to integrate seamlessly into human environments. Critical to their coexistence and cooperation with humans is the ability to understand natural language communications and exhibit human-like behaviors. This work focuses on generating diverse whole-body motions for humanoid robots from language descriptions. We leverage human motion priors from extensive human motion datasets to initialize humanoid motions and employ the commonsense reasoning capabilities of Vision Language Models (VLMs) to edit and refine these motions. Our approach demonstrates the capability to produce natural, expressive, and text-aligned humanoid motions, validated through both simulated and real-world experiments. More videos can be found at https://ut-austin-rpl.github.io/Harmon/., Comment: Accepted for oral presentation at 8th Annual Conference on Robot Learning. Project website: https://ut-austin-rpl.github.io/Harmon/

The geometric shape, symmetry, and topology of colloidal particles often allow for controlling colloidal phase behavior and physical properties of these soft matter systems. In liquid crystalline dispersions, colloidal particles with low symmetry and nontrivial topology of surface confine-ment are of particular interest, including surfaces shaped as handlebodies, spirals, knots, mul-ti-component links, and so on. These types of colloidal surfaces induce topologically nontrivial three-dimensional director field configurations and topological defects. Director switching by electric fields, laser tweezing of defects, and local photo-thermal melting of the liquid crystal host medium promote transformations among many stable and metastable particle-induced director configurations that can be revealed by means of direct label-free three-dimensional nonlinear op-tical imaging. The interplay between topologies of colloidal surfaces, director fields, and defects is found to show a number of unexpected features, such as knotting and linking of line defects, often uniquely arising from the nonpolar nature of the nematic director field. This review article high-lights fascinating examples of new physical behavior arising from the interplay of nematic mo-lecular order and both chiral symmetry and topology of colloidal inclusions within the nematic host. Furthermore, the article concludes with a brief discussion of how these findings may lay the groundwork for new types of topology-dictated self-assembly in soft condensed matter leading to novel mesostructured composite materials, as well as for experimental insights into the pure-math aspects of low-dimensional topology., Comment: 16 pages, 5 figures

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

Low-rank training has emerged as a promising approach for reducing memory usage in training Large Language Models (LLMs). Previous methods either rely on decomposing weight matrices (e.g., LoRA), or seek to decompose gradient matrices (e.g., GaLore) to ensure reduced memory consumption. However, both of them constrain the training in a low-rank subspace, thus inevitably leading to sub-optimal performance. This raises a question: whether it is possible to consistently preserve the low-rank constraint for memory efficiency, while achieving full-rank training (i.e., training with full-rank gradients of full-rank weights) to avoid inferior outcomes? In this paper, we propose a new plug-and-play training framework for LLMs called Fira, as the first attempt to achieve this goal. First, we observe an interesting phenomenon during LLM training: the scaling impact of adaptive optimizers (e.g., Adam) on the gradient norm remains similar from low-rank to full-rank training. Based on this observation, we propose a norm-based scaling method, which utilizes the scaling impact of low-rank optimizers as substitutes for that of original full-rank optimizers to enable full-rank training. In this way, we can preserve the low-rank constraint in the optimizer while achieving full-rank training for better performance. Moreover, we find that there are sudden gradient rises during the optimization process, potentially causing loss spikes. To address this, we further put forward a norm-growth limiter to smooth the gradient via regulating the relative increase of gradient norms. Extensive experiments on the pre-training and fine-tuning of LLMs show that Fira outperforms both LoRA and GaLore, achieving performance that is comparable to or even better than full-rank training., Comment: Add further analysis of the scaling factor, code is available at: https://github.com/xichen-fy/Fira

Using high-resolution x-ray tomography, we experimentally investigate the bridge structures in tapped granular packings composed of particles with varying friction coefficients. We find that gravity can induce subtle structural changes on the load-bearing contacts, allowing us to identify the correct load-bearing contacts based on structural information alone. Using these identified load-bearing contacts, we investigate the cooperative bridge structures which are mechanical backbones of the system. We characterize the geometric properties of these bridges and find that their cooperativity increases as the packing fraction decreases. The knowledge of bridges can enhance our understanding of the rheological properties of granular materials., Comment: 21 pages, 5 figures

Data selection is of great significance in pre-training large language models, given the variation in quality within the large-scale available training corpora. To achieve this, researchers are currently investigating the use of data influence to measure the importance of data instances, $i.e.,$ a high influence score indicates that incorporating this instance to the training set is likely to enhance the model performance. Consequently, they select the top-$k$ instances with the highest scores. However, this approach has several limitations. (1) Computing the influence of all available data is time-consuming. (2) The selected data instances are not diverse enough, which may hinder the pre-trained model's ability to generalize effectively to various downstream tasks. In this paper, we introduce \texttt{Quad}, a data selection approach that considers both quality and diversity by using data influence to achieve state-of-the-art pre-training results. In particular, noting that attention layers capture extensive semantic details, we have adapted the accelerated $iHVP$ computation methods for attention layers, enhancing our ability to evaluate the influence of data, $i.e.,$ its quality. For the diversity, \texttt{Quad} clusters the dataset into similar data instances within each cluster and diverse instances across different clusters. For each cluster, if we opt to select data from it, we take some samples to evaluate the influence to prevent processing all instances. To determine which clusters to select, we utilize the classic Multi-Armed Bandit method, treating each cluster as an arm. This approach favors clusters with highly influential instances (ensuring high quality) or clusters that have been selected less frequently (ensuring diversity), thereby well balancing between quality and diversity.

A second-order-based latent factor (SLF) analysis model demonstrates superior performance in graph representation learning, particularly for high-dimensional and incomplete (HDI) interaction data, by incorporating the curvature information of the loss landscape. However, its objective function is commonly bi-linear and non-convex, causing the SLF model to suffer from a low convergence rate. To address this issue, this paper proposes a PID controller-incorporated SLF (PSLF) model, leveraging two key strategies: a) refining learning error estimation by incorporating the PID controller principles, and b) acquiring second-order information insights through Hessian-vector products. Experimental results on multiple HDI datasets indicate that the proposed PSLF model outperforms four state-of-the-art latent factor models based on advanced optimizers regarding convergence rates and generalization performance.

Estimating global human motion from moving cameras is challenging due to the entanglement of human and camera motions. To mitigate the ambiguity, existing methods leverage learned human motion priors, which however often result in oversmoothed motions with misaligned 2D projections. To tackle this problem, we propose COIN, a control-inpainting motion diffusion prior that enables fine-grained control to disentangle human and camera motions. Although pre-trained motion diffusion models encode rich motion priors, we find it non-trivial to leverage such knowledge to guide global motion estimation from RGB videos. COIN introduces a novel control-inpainting score distillation sampling method to ensure well-aligned, consistent, and high-quality motion from the diffusion prior within a joint optimization framework. Furthermore, we introduce a new human-scene relation loss to alleviate the scale ambiguity by enforcing consistency among the humans, camera, and scene. Experiments on three challenging benchmarks demonstrate the effectiveness of COIN, which outperforms the state-of-the-art methods in terms of global human motion estimation and camera motion estimation. As an illustrative example, COIN outperforms the state-of-the-art method by 33% in world joint position error (W-MPJPE) on the RICH dataset., Comment: ECCV 2024

Image segmentation is a long-standing challenge in computer vision, studied continuously over several decades, as evidenced by seminal algorithms such as N-Cut, FCN, and MaskFormer. With the advent of foundation models (FMs), contemporary segmentation methodologies have embarked on a new epoch by either adapting FMs (e.g., CLIP, Stable Diffusion, DINO) for image segmentation or developing dedicated segmentation foundation models (e.g., SAM). These approaches not only deliver superior segmentation performance, but also herald newfound segmentation capabilities previously unseen in deep learning context. However, current research in image segmentation lacks a detailed analysis of distinct characteristics, challenges, and solutions associated with these advancements. This survey seeks to fill this gap by providing a thorough review of cutting-edge research centered around FM-driven image segmentation. We investigate two basic lines of research -- generic image segmentation (i.e., semantic segmentation, instance segmentation, panoptic segmentation), and promptable image segmentation (i.e., interactive segmentation, referring segmentation, few-shot segmentation) -- by delineating their respective task settings, background concepts, and key challenges. Furthermore, we provide insights into the emergence of segmentation knowledge from FMs like CLIP, Stable Diffusion, and DINO. An exhaustive overview of over 300 segmentation approaches is provided to encapsulate the breadth of current research efforts. Subsequently, we engage in a discussion of open issues and potential avenues for future research. We envisage that this fresh, comprehensive, and systematic survey catalyzes the evolution of advanced image segmentation systems. A public website is created to continuously track developments in this fast advancing field: \url{https://github.com/stanley-313/ImageSegFM-Survey}., Comment: A comprehensive survey of image segmentation in foundation model era

Random feature attention (RFA) adopts random fourier feature (RFF) methods to approximate the softmax function, resulting in a linear time and space attention mechanism that enables the construction of an efficient Transformer. Inspired by RFA, we propose Macformer, a Transformer architecture that employs random Maclaurin features (RMF) to approximate various dot-product kernels, thereby accelerating attention computations for long sequence. Macformer consists of Random Maclaurin Feature Attention (RMFA) and pre-post Scaling Batch Normalization (ppSBN), the former is an unbiased approximation for dot-product kernelized attention and the later is a two-stage regularization mechanism guaranteeing the error of RMFA. We conducted toy experiments to demonstrate the efficiency of RMFA and ppSBN, and experiments on long range arena (LRA) benchmark to validate the acceleration and accuracy of Macformer with different dot-product kernels. Experiment results of Macformer are consistent with our theoretical analysis.

Federated Learning (FL) has emerged as an excellent solution for performing deep learning on different data owners without exchanging raw data. However, statistical heterogeneity in FL presents a key challenge, leading to a phenomenon of skewness in local model parameter distributions that researchers have largely overlooked. In this work, we propose the concept of parameter skew to describe the phenomenon that can substantially affect the accuracy of global model parameter estimation. Additionally, we introduce FedSA, an aggregation strategy to obtain a high-quality global model, to address the implication from parameter skew. Specifically, we categorize parameters into high-dispersion and low-dispersion groups based on the coefficient of variation. For high-dispersion parameters, Micro-Classes (MIC) and Macro-Classes (MAC) represent the dispersion at the micro and macro levels, respectively, forming the foundation of FedSA. To evaluate the effectiveness of FedSA, we conduct extensive experiments with different FL algorithms on three computer vision datasets. FedSA outperforms eight state-of-the-art baselines by about 4.7% in test accuracy.

Graph neural networks (GNNs) have garnered significant attention due to their ability to represent graph data. Among various GNN variants, graph attention network (GAT) stands out since it is able to dynamically learn the importance of different nodes. However, present GATs heavily rely on the smoothed node features to obtain the attention coefficients rather than graph structural information, which fails to provide crucial contextual cues for node representations. To address this issue, this study proposes a neighbor overlay-induced graph attention network (NO-GAT) with the following two-fold ideas: a) learning favorable structural information, i.e., overlaid neighbors, outside the node feature propagation process from an adjacency matrix; b) injecting the information of overlaid neighbors into the node feature propagation process to compute the attention coefficient jointly. Empirical studies on graph benchmark datasets indicate that the proposed NO-GAT consistently outperforms state-of-the-art models.

In recent years, with the rapid application of large language models across various fields, the scale of these models has gradually increased, and the resources required for their pre-training have grown exponentially. Training an LLM from scratch will cost a lot of computation resources while scaling up from a smaller model is a more efficient approach and has thus attracted significant attention. In this paper, we present AquilaMoE, a cutting-edge bilingual 8*16B Mixture of Experts (MoE) language model that has 8 experts with 16 billion parameters each and is developed using an innovative training methodology called EfficientScale. This approach optimizes performance while minimizing data requirements through a two-stage process. The first stage, termed Scale-Up, initializes the larger model with weights from a pre-trained smaller model, enabling substantial knowledge transfer and continuous pretraining with significantly less data. The second stage, Scale-Out, uses a pre-trained dense model to initialize the MoE experts, further enhancing knowledge transfer and performance. Extensive validation experiments on 1.8B and 7B models compared various initialization schemes, achieving models that maintain and reduce loss during continuous pretraining. Utilizing the optimal scheme, we successfully trained a 16B model and subsequently the 8*16B AquilaMoE model, demonstrating significant improvements in performance and training efficiency.

Retrieval-augmented generation (RAG) is a framework enabling large language models (LLMs) to enhance their accuracy and reduce hallucinations by integrating external knowledge bases. In this paper, we introduce a hybrid RAG system enhanced through a comprehensive suite of optimizations that significantly improve retrieval quality, augment reasoning capabilities, and refine numerical computation ability. We refined the text chunks and tables in web pages, added attribute predictors to reduce hallucinations, conducted LLM Knowledge Extractor and Knowledge Graph Extractor, and finally built a reasoning strategy with all the references. We evaluated our system on the CRAG dataset through the Meta CRAG KDD Cup 2024 Competition. Both the local and online evaluations demonstrate that our system significantly enhances complex reasoning capabilities. In local evaluations, we have significantly improved accuracy and reduced error rates compared to the baseline model, achieving a notable increase in scores. In the meanwhile, we have attained outstanding results in online assessments, demonstrating the performance and generalization capabilities of the proposed system. The source code for our system is released in \url{https://gitlab.aicrowd.com/shizueyy/crag-new}., Comment: Technical report for 3rd prize in Task 1 of Meta CRAG KDD Cup 2024

High quality spatiotemporal signal is vitally important for real application scenarios like energy management, traffic planning and cyber security. Due to the uncontrollable factors like abrupt sensors breakdown or communication fault, the spatiotemporal signal collected by sensors is always incomplete. A dynamic graph convolutional network (DGCN) is effective for processing spatiotemporal signal recovery. However, it adopts a static GCN and a sequence neural network to explore the spatial and temporal patterns, separately. Such a separated two-step processing is loose spatiotemporal, thereby failing to capture the complex inner spatiotemporal correlation. To address this issue, this paper proposes a Compact-Dynamic Graph Convolutional Network (CDGCN) for spatiotemporal signal recovery with the following two-fold ideas: a) leveraging the tensor M-product to build a unified tensor graph convolution framework, which considers both spatial and temporal patterns simultaneously; and b) constructing a differential smoothness-based objective function to reduce the noise interference in spatiotemporal signal, thereby further improve the recovery accuracy. Experiments on real-world spatiotemporal datasets demonstrate that the proposed CDGCN significantly outperforms the state-of-the-art models in terms of recovery accuracy.

Generative artificial intelligence (GAI) is a promising technique towards 6G networks, and generative foundation models such as large language models (LLMs) have attracted considerable interest from academia and telecom industry. This work considers a novel edge-cloud deployment of foundation models in 6G networks. Specifically, it aims to minimize the service delay of foundation models by radio resource allocation and task offloading, i.e., offloading diverse content generation tasks to proper LLMs at the network edge or cloud. In particular, we first introduce the communication system model, i.e., allocating radio resources and calculating link capacity to support generated content transmission, and then we present the LLM inference model to calculate the delay of content generation. After that, we propose a novel in-context learning method to optimize the task offloading decisions. It utilizes LLM's inference capabilities, and avoids the difficulty of dedicated model training or fine-tuning as in conventional machine learning algorithms. Finally, the simulations demonstrate that the proposed edge-cloud deployment and in-context learning task offloading method can achieve satisfactory generation service quality without dedicated model training or fine-tuning.

A dynamic weighted directed graph (DWDG) is commonly encountered in various application scenarios. It involves extensive dynamic interactions among numerous nodes. Most existing approaches explore the intricate temporal patterns hidden in a DWDG from the purely data-driven perspective, which suffers from accuracy loss when a DWDG exhibits strong fluctuations over time. To address this issue, this study proposes a novel Extended-Kalman-Filter-Incorporated Latent Feature (EKLF) model to represent a DWDG from the model-driven perspective. Its main idea is divided into the following two-fold ideas: a) adopting a control model, i.e., the Extended Kalman Filter (EKF), to track the complex temporal patterns precisely with its nonlinear state-transition and observation functions; and b) introducing an alternating least squares (ALS) algorithm to train the latent features (LFs) alternatively for precisely representing a DWDG. Empirical studies on DWDG datasets demonstrate that the proposed EKLF model outperforms state-of-the-art models in prediction accuracy and computational efficiency for missing edge weights of a DWDG. It unveils the potential for precisely representing a DWDG by incorporating a control model.

Large language model (LLM) has recently been considered a promising technique for many fields. This work explores LLM-based wireless network optimization via in-context learning. To showcase the potential of LLM technologies, we consider the base station (BS) power control as a case study, a fundamental but crucial technique that is widely investigated in wireless networks. Different from existing machine learning (ML) methods, our proposed in-context learning algorithm relies on LLM's inference capabilities. It avoids the complexity of tedious model training and hyper-parameter fine-tuning, which is a well-known bottleneck of many ML algorithms. Specifically, the proposed algorithm first describes the target task via formatted natural language, and then designs the in-context learning framework and demonstration examples. After that, it considers two cases, namely discrete-state and continuous-state problems, and proposes state-based and ranking-based methods to select appropriate examples for these two cases, respectively. Finally, the simulations demonstrate that the proposed algorithm can achieve comparable performance as conventional deep reinforcement learning (DRL) techniques without dedicated model training or fine-tuning. Such an efficient and low-complexity approach has great potential for future wireless network optimization.

Large language models (LLMs) have demonstrated great success in various fields, benefiting from their huge amount of parameters that store knowledge. However, LLMs still suffer from several key issues, such as hallucination problems, knowledge update issues, and lacking domain-specific expertise. The appearance of retrieval-augmented generation (RAG), which leverages an external knowledge database to augment LLMs, makes up those drawbacks of LLMs. This paper reviews all significant techniques of RAG, especially in the retriever and the retrieval fusions. Besides, tutorial codes are provided for implementing the representative techniques in RAG. This paper further discusses the RAG training, including RAG with/without datastore update. Then, we introduce the application of RAG in representative natural language processing tasks and industrial scenarios. Finally, this paper discusses the future directions and challenges of RAG for promoting its development.

The availability of massive vehicle trajectory data enables the modeling of road-network constrained movement as travel-cost distributions rather than just single-valued costs, thereby capturing the inherent uncertainty of movement and enabling improved routing quality. Thus, stochastic routing has been studied extensively in the edge-centric model, where such costs are assigned to the edges in a graph representation of a road network. However, as this model still disregards important information in trajectories and fails to capture dependencies among cost distributions, a path-centric model, where costs are assigned to paths, has been proposed that captures dependencies better and provides an improved foundation for routing. Unfortunately, when applied in this model, existing routing algorithms are inefficient due to two shortcomings that we eliminate. First, when exploring candidate paths, existing algorithms only consider the costs of candidate paths from the source to intermediate vertices, while disregarding the costs of travel from the intermediate vertices to the destination, causing many non-competitive paths to be explored. We propose two heuristics for estimating the cost from an intermediate vertex to the destination, thus improving routing efficiency. Second, the edge-centric model relies on stochastic dominance-based pruning to improve efficiency. This pruning assumes that costs are independent and is therefore inapplicable in the path-centric model that takes dependencies into account. We introduce a notion of virtual path that effectively enables stochastic dominance-based pruning in the path-based model, thus further improving efficiency. Empirical studies using two real-world trajectory sets offer insight into the properties of the proposed solution, indicating that it enables efficient stochastic routing in the path-centric model.

Visually impaired people are a large group who can only use braille for reading and writing. However, the lack of special educational resources is the bottleneck for educating them. Educational equity is a reflection of the level of social civilization, cultural equality, and individual dignity. Facilitating and improving lifelong learning channels for the visually impaired is of great significance. Their written braille homework or exam papers cannot be understood by sighted teachers, because of the lack of a highly accurate braille translation system, especially in Chinese which has tone marks. braille writers often omit tone marks to save space, leading to confusion when braille with the same consonants and vowels is translated into Chinese. Previous algorithms were insufficient in extracting contextual information, resulting in low accuracy of braille translations into Chinese. This project informatively fine-tuned the mT5 model with an Encoder-decoder architecture for braille to Chinese character conversion. This research created a training set of braille and corresponding Chinese text from the Leipzig Corpora. This project significantly reduced the confusion in braille, achieving $62.4$ and $62.3$ BLEU scores in the validation and test sets, with a curriculum learning fine-tuning method. By incorporating the braille recognition algorithm, this project is the first publicly available braille translation system and can benefit lots of visually impaired students and families who are preparing for the Chinese College Test and help to propel their college dreams in the future. There is a demo on our homepage\footnote{\url{https://vision-braille.com/}}., Comment: This paper is submitted to NeurIPS 2024 High School Project Track

Large Multimodal Models (LMMs) exhibit impressive cross-modal understanding and reasoning abilities, often assessed through multiple-choice questions (MCQs) that include an image, a question, and several options. However, many benchmarks used for such evaluations suffer from systematic biases. Remarkably, Large Language Models (LLMs) without any visual perception capabilities achieve non-trivial performance, undermining the credibility of these evaluations. To address this issue while maintaining the efficiency of MCQ evaluations, we propose MMEvalPro, a benchmark designed to avoid Type-I errors through a trilogy evaluation pipeline and more rigorous metrics. For each original question from existing benchmarks, human annotators augment it by creating one perception question and one knowledge anchor question through a meticulous annotation process. MMEvalPro comprises $2,138$ question triplets, totaling $6,414$ distinct questions. Two-thirds of these questions are manually labeled by human experts, while the rest are sourced from existing benchmarks (MMMU, ScienceQA, and MathVista). Compared with the existing benchmarks, our experiments with the latest LLMs and LMMs demonstrate that MMEvalPro is more challenging (the best LMM lags behind human performance by $31.73\%$, compared to an average gap of $8.03\%$ in previous benchmarks) and more trustworthy (the best LLM trails the best LMM by $23.09\%$, whereas the gap for previous benchmarks is just $14.64\%$). Our in-depth analysis explains the reason for the large performance gap and justifies the trustworthiness of evaluation, underscoring its significant potential for advancing future research., Comment: 21 pages, code released at https://github.com/chenllliang/MMEvalPro, Homepage at https://mmevalpro.github.io/

We present SMPLOlympics, a collection of physically simulated environments that allow humanoids to compete in a variety of Olympic sports. Sports simulation offers a rich and standardized testing ground for evaluating and improving the capabilities of learning algorithms due to the diversity and physically demanding nature of athletic activities. As humans have been competing in these sports for many years, there is also a plethora of existing knowledge on the preferred strategy to achieve better performance. To leverage these existing human demonstrations from videos and motion capture, we design our humanoid to be compatible with the widely-used SMPL and SMPL-X human models from the vision and graphics community. We provide a suite of individual sports environments, including golf, javelin throw, high jump, long jump, and hurdling, as well as competitive sports, including both 1v1 and 2v2 games such as table tennis, tennis, fencing, boxing, soccer, and basketball. Our analysis shows that combining strong motion priors with simple rewards can result in human-like behavior in various sports. By providing a unified sports benchmark and baseline implementation of state and reward designs, we hope that SMPLOlympics can help the control and animation communities achieve human-like and performant behaviors., Comment: Project page: https://smplolympics.github.io/SMPLOlympics

The emerging video LMMs (Large Multimodal Models) have achieved significant improvements on generic video understanding in the form of VQA (Visual Question Answering), where the raw videos are captured by cameras. However, a large portion of videos in real-world applications are edited videos, \textit{e.g.}, users usually cut and add effects/modifications to the raw video before publishing it on social media platforms. The edited videos usually have high view counts but they are not covered in existing benchmarks of video LMMs, \textit{i.e.}, ActivityNet-QA, or VideoChatGPT benchmark. In this paper, we leverage the edited videos on a popular short video platform, \textit{i.e.}, TikTok, and build a video VQA benchmark (named EditVid-QA) covering four typical editing categories, i.e., effect, funny, meme, and game. Funny and meme videos benchmark nuanced understanding and high-level reasoning, while effect and game evaluate the understanding capability of artificial design. Most of the open-source video LMMs perform poorly on the EditVid-QA benchmark, indicating a huge domain gap between edited short videos on social media and regular raw videos. To improve the generalization ability of LMMs, we collect a training set for the proposed benchmark based on both Panda-70M/WebVid raw videos and small-scale TikTok/CapCut edited videos, which boosts the performance on the proposed EditVid-QA benchmark, indicating the effectiveness of high-quality training data. We also identified a serious issue in the existing evaluation protocol using the GPT-3.5 judge, namely a "sorry" attack, where a sorry-style naive answer can achieve an extremely high rating from the GPT judge, e.g., over 4.3 for correctness score on VideoChatGPT evaluation protocol. To avoid the "sorry" attacks, we evaluate results with GPT-4 judge and keyword filtering. The dataset is released at https://github.com/XenonLamb/EditVid-QA.

Given a hypergraph, influence maximization (IM) is to discover a seed set containing $k$ vertices that have the maximal influence. Although the existing vertex-based IM algorithms perform better than the hyperedge-based algorithms by generating random reverse researchable (RR) sets, they are inefficient because (i) they ignore important structural information associated with hyperedges and thus obtain inferior results, (ii) the frequently-used sampling methods for generating RR sets have low efficiency because of a large number of required samplings along with high sampling variances, and (iii) the vertex-based IM algorithms have large overheads in terms of running time and memory costs. To overcome these shortcomings, this paper proposes a novel approach, called \emph{HyperIM}. The key idea behind \emph{HyperIM} is to differentiate structural information of vertices for developing stratified sampling combined with highly-efficient strategies to generate the RR sets. With theoretical guarantees, \emph{HyperIM} is able to accelerate the influence spread, improve the sampling efficiency, and cut down the expected running time. To further reduce the running time and memory costs, we optimize \emph{HyperIM} by inferring the bound of the required number of RR sets in conjunction with stratified sampling. Experimental results on real-world hypergraphs show that \emph{HyperIM} is able to reduce the number of required RR sets and running time by orders of magnitude while increasing the influence spread by up to $2.73X$ on average, compared to the state-of-the-art IM algorithms., Comment: 15 pages,10figures

Chain-of-thought distillation is a powerful technique for transferring reasoning abilities from large language models (LLMs) to smaller student models. Previous methods typically require the student to mimic the step-by-step rationale produced by LLMs, often facing the following challenges: (i) Tokens within a rationale vary in significance, and treating them equally may fail to accurately mimic keypoint tokens, leading to reasoning errors. (ii) They usually distill knowledge by consistently predicting all the steps in a rationale, which falls short in distinguishing the learning order of step generation. This diverges from the human cognitive progression of starting with easy tasks and advancing to harder ones, resulting in sub-optimal outcomes. To this end, we propose a unified framework, called KPOD, to address these issues. Specifically, we propose a token weighting module utilizing mask learning to encourage accurate mimicry of keypoint tokens by the student during distillation. Besides, we develop an in-rationale progressive distillation strategy, starting with training the student to generate the final reasoning steps and gradually extending to cover the entire rationale. To accomplish this, a weighted token generation loss is proposed to assess step reasoning difficulty, and a value function is devised to schedule the progressive distillation by considering both step difficulty and question diversity. Extensive experiments on four reasoning benchmarks illustrate our KPOD outperforms previous methods by a large margin., Comment: Accepted by ICML 2024

Offline model-based optimization (MBO) aims to maximize a black-box objective function using only an offline dataset of designs and scores. These tasks span various domains, such as robotics, material design, and protein and molecular engineering. A common approach involves training a surrogate model using existing designs and their corresponding scores, and then generating new designs through gradient-based updates with respect to the surrogate model. This method suffers from the out-of-distribution issue, where the surrogate model may erroneously predict high scores for unseen designs. To address this challenge, we introduce a novel method, Design Editing for Offline Model-based Optimization} (DEMO), which leverages a diffusion prior to calibrate overly optimized designs. DEMO first generates pseudo design candidates by performing gradient ascent with respect to a surrogate model. Then, an editing process refines these pseudo design candidates by introducing noise and subsequently denoising them with a diffusion prior trained on the offline dataset, ensuring they align with the distribution of valid designs. We provide a theoretical proof that the difference between the final optimized designs generated by DEMO and the prior distribution of the offline dataset is controlled by the noise injected during the editing process. Empirical evaluations on seven offline MBO tasks show that DEMO outperforms various baseline methods, achieving the highest mean rank of 2.1 and a median rank of 1.

Large language models (LLMs) have received considerable attention recently due to their outstanding comprehension and reasoning capabilities, leading to great progress in many fields. The advancement of LLM techniques also offers promising opportunities to automate many tasks in the telecommunication (telecom) field. After pre-training and fine-tuning, LLMs can perform diverse downstream tasks based on human instructions, paving the way to artificial general intelligence (AGI)-enabled 6G. Given the great potential of LLM technologies, this work aims to provide a comprehensive overview of LLM-enabled telecom networks. In particular, we first present LLM fundamentals, including model architecture, pre-training, fine-tuning, inference and utilization, model evaluation, and telecom deployment. Then, we introduce LLM-enabled key techniques and telecom applications in terms of generation, classification, optimization, and prediction problems. Specifically, the LLM-enabled generation applications include telecom domain knowledge, code, and network configuration generation. After that, the LLM-based classification applications involve network security, text, image, and traffic classification problems. Moreover, multiple LLM-enabled optimization techniques are introduced, such as automated reward function design for reinforcement learning and verbal reinforcement learning. Furthermore, for LLM-aided prediction problems, we discussed time-series prediction models and multi-modality prediction problems for telecom. Finally, we highlight the challenges and identify the future directions of LLM-enabled telecom networks.

Counting small cohesive subgraphs in a graph is a fundamental operation with numerous applications in graph analysis. Previous studies on cohesive subgraph counting are mainly based on the clique model, which aim to count the number of $k$-cliques in a graph with a small $k$. However, the clique model often proves too restrictive for practical use. To address this issue, we investigate a new problem of counting cohesive subgraphs that adhere to the hereditary property. Here the hereditary property means that if a graph $G$ has a property $\mathcal{P}$, then any induced subgraph of $G$ also has a property $\mathcal{P}$. To count these hereditary cohesive subgraphs (\hcss), we propose a new listing-based framework called \hcslist, which employs a backtracking enumeration procedure to count all \hcss. A notable limitation of \hcslist is that it requires enumerating all \hcss, making it intractable for large and dense graphs due to the exponential growth in the number of \hcss with respect to graph size. To overcome this limitation, we propose a novel pivot-based framework called \hcspivot, which can count most \hcss in a combinatorial manner without explicitly listing them. Two additional noteworthy features of \hcspivot is its ability to (1) simultaneously count \hcss of any size and (2) simultaneously count \hcss for each vertex or each edge, while \hcslist is only capable of counting a specific size of \hcs and obtaining a total count of \hcss in a graph. We focus specifically on two \hcs: $s$-defective clique and $s$-plex, with several non-trivial pruning techniques to enhance the efficiency. We conduct extensive experiments on 8 large real-world graphs, and the results demonstrate the high efficiency and effectiveness of our solutions.

Benchmarking is crucial for evaluating a DBMS, yet existing benchmarks often fail to reflect the varied nature of user workloads. As a result, there is increasing momentum toward creating databases that incorporate real-world user data to more accurately mirror business environments. However, privacy concerns deter users from directly sharing their data, underscoring the importance of creating synthesized databases for benchmarking that also prioritize privacy protection. Differential privacy has become a key method for safeguarding privacy when sharing data, but the focus has largely been on minimizing errors in aggregate queries or classification tasks, with less attention given to benchmarking factors like runtime performance. This paper delves into the creation of privacy-preserving databases specifically for benchmarking, aiming to produce a differentially private database whose query performance closely resembles that of the original data. Introducing PrivBench, an innovative synthesis framework, we support the generation of high-quality data that maintains privacy. PrivBench uses sum-product networks (SPNs) to partition and sample data, enhancing data representation while securing privacy. The framework allows users to adjust the detail of SPN partitions and privacy settings, crucial for customizing privacy levels. We validate our approach, which uses the Laplace and exponential mechanisms, in maintaining privacy. Our tests show that PrivBench effectively generates data that maintains privacy and excels in query performance, consistently reducing errors in query execution time, query cardinality, and KL divergence.

We address the challenge of content diversity and controllability in pedestrian simulation for driving scenarios. Recent pedestrian animation frameworks have a significant limitation wherein they primarily focus on either following trajectory [46] or the content of the reference video [57], consequently overlooking the potential diversity of human motion within such scenarios. This limitation restricts the ability to generate pedestrian behaviors that exhibit a wider range of variations and realistic motions and therefore restricts its usage to provide rich motion content for other components in the driving simulation system, e.g., suddenly changed motion to which the autonomous vehicle should respond. In our approach, we strive to surpass the limitation by showcasing diverse human motions obtained from various sources, such as generated human motions, in addition to following the given trajectory. The fundamental contribution of our framework lies in combining the motion tracking task with trajectory following, which enables the tracking of specific motion parts (e.g., upper body) while simultaneously following the given trajectory by a single policy. This way, we significantly enhance both the diversity of simulated human motion within the given scenario and the controllability of the content, including language-based control. Our framework facilitates the generation of a wide range of human motions, contributing to greater realism and adaptability in pedestrian simulations for driving scenarios. More information is on our project page https://wangjingbo1219.github.io/papers/CVPR2024_PACER_PLUS/PACERPLUSPage.html .

We present a new challenge to examine whether large language models understand social norms. In contrast to existing datasets, our dataset requires a fundamental understanding of social norms to solve. Our dataset features the largest set of social norm skills, consisting of 402 skills and 12,383 questions covering a wide set of social norms ranging from opinions and arguments to culture and laws. We design our dataset according to the K-12 curriculum. This enables the direct comparison of the social understanding of large language models to humans, more specifically, elementary students. While prior work generates nearly random accuracy on our benchmark, recent large language models such as GPT3.5-Turbo and LLaMA2-Chat are able to improve the performance significantly, only slightly below human performance. We then propose a multi-agent framework based on large language models to improve the models' ability to understand social norms. This method further improves large language models to be on par with humans. Given the increasing adoption of large language models in real-world applications, our finding is particularly important and presents a unique direction for future improvements.

We proposed an end-to-end system design towards utilizing Retrieval Augmented Generation (RAG) to improve the factual accuracy of Large Language Models (LLMs) for domain-specific and time-sensitive queries related to private knowledge-bases. Our system integrates RAG pipeline with upstream datasets processing and downstream performance evaluation. Addressing the challenge of LLM hallucinations, we finetune models with a curated dataset which originates from CMU's extensive resources and annotated with the teacher model. Our experiments demonstrate the system's effectiveness in generating more accurate answers to domain-specific and time-sensitive inquiries. The results also revealed the limitations of fine-tuning LLMs with small-scale and skewed datasets. This research highlights the potential of RAG systems in augmenting LLMs with external datasets for improved performance in knowledge-intensive tasks. Our code and models are available on Github., Comment: These authors contributed equally to this work