Search

Large language models (LLM) have achieved remarkable success in natural language generation but lesser focus has been given to their applicability in decision making tasks such as classification. We show that LLMs like LLaMa can achieve high performance on large multi-class classification tasks but still make classification errors and worse, generate out-of-vocabulary class labels. To address these critical issues, we introduce Paraphrase and AGgregate (PAG)-LLM approach wherein an LLM generates multiple paraphrases of the input query (parallel queries), performs multi-class classification for the original query and each paraphrase, and at the end aggregate all the classification labels based on their confidence scores. We evaluate PAG-LLM on two large multi-class classication datasets: CLINC, and Banking and show 22.7% and 15.1% error reduction. We show that PAG-LLM is especially effective for hard examples where LLM is uncertain, and reduces the critical misclassification and hallucinated label generation errors, Comment: Accepted at SIGIR 2024

Confinement/deconfinement, captivating attributes of high-energy elementary particles, have recently garnered wide attention in quantum simulations based on cold atoms. Yet, the partial confinement, an intermediate state between the confinement and deconfinement, remains underexplored. The partial confinement encapsulates the phenomenon that the confining behavior of charged particles is contingent upon their relative positions. In this paper, we demonstrate that the spin-1 quantum link model provides an excellent platform for exploring partial confinement. We conduct a comprehensive investigation of the physics emerging from partial confinement in both the context of equilibrium and non-equilibrium dynamics. Potential experimental setups using cold atoms are also discussed. Our work offers a simple and feasible routine for the study of confinement-related physics in the state-of-the-art artificial quantum systems subject to gauge symmetries.

The eighth AI City Challenge highlighted the convergence of computer vision and artificial intelligence in areas like retail, warehouse settings, and Intelligent Traffic Systems (ITS), presenting significant research opportunities. The 2024 edition featured five tracks, attracting unprecedented interest from 726 teams in 47 countries and regions. Track 1 dealt with multi-target multi-camera (MTMC) people tracking, highlighting significant enhancements in camera count, character number, 3D annotation, and camera matrices, alongside new rules for 3D tracking and online tracking algorithm encouragement. Track 2 introduced dense video captioning for traffic safety, focusing on pedestrian accidents using multi-camera feeds to improve insights for insurance and prevention. Track 3 required teams to classify driver actions in a naturalistic driving analysis. Track 4 explored fish-eye camera analytics using the FishEye8K dataset. Track 5 focused on motorcycle helmet rule violation detection. The challenge utilized two leaderboards to showcase methods, with participants setting new benchmarks, some surpassing existing state-of-the-art achievements., Comment: Summary of the 8th AI City Challenge Workshop in conjunction with CVPR 2024

From the mid-19th century to the end of the 20th century, photographic plates served as the primary detectors for astronomical observations. Astronomical photographic observations in China began in 1901, and over a century, a total of approximately 30,000 astronomical photographic plates have been captured. These historical plates play an irreplaceable role in conducting long-term, time-domain astronomical research. To preserve and explore these valuable original astronomical observational data, Shanghai Astronomical Observatory has organized the transportation of plates taken at night from various stations across the country to the Sheshan Plate Archive for centralized preservation. For the first time, plate information statistics was performed. On this basis, the plates were cleaned and digitally scanned, and finally digitized images were acquired for 29,314 plates. In this study, using Gaia DR2 as the reference star catalog, astrometric processing has been carried out successfully on 15,696 single-exposure plates, including object extraction, stellar identification, and plate model computation. As a result, for long focal length telescopes, such as the 40cm double-tube refractor telescope and the 1.56m reflector telescope at the Shanghai Astronomical Observatory and the 1m reflector telescope at the Yunnan Astronomical Observatory, the astrometric accuracy obtained for their plates is approximately 0.1" to 0.3". The distribution of astrometric accuracy for medium and short focal length telescopes ranges from 0.3" to 1.0". The relevant data of this batch of plates, including digitized images and stellar catalog of the plates are archived and released by the National Astronomical Data Center. Users can access and download plate data based on keywords such as station, telescope, observation year, and observed celestial coordinates., Comment: Accepted for Research in Astronomy and Astrophysics, 17 pages, 14 figures, 6 tables. Database, https://nadc.china-vo.org/res/r100742/

Automating the checkout process is important in smart retail, where users effortlessly pass products by hand through a camera, triggering automatic product detection, tracking, and counting. In this emerging area, due to the lack of annotated training data, we introduce a dataset comprised of product 3D models, which allows for fast, flexible, and large-scale training data generation through graphic engine rendering. Within this context, we discern an intriguing facet, because of the user "hands-on" approach, bias in user behavior leads to distinct patterns in the real checkout process. The existence of such patterns would compromise training effectiveness if training data fail to reflect the same. To address this user bias problem, we propose a training data optimization framework, i.e., training with digital twins (DtTrain). Specifically, we leverage the product 3D models and optimize their rendering viewpoint and illumination to generate "digital twins" that visually resemble representative user images. These digital twins, inherit product labels and, when augmented, form the Digital Twin training set (DT set). Because the digital twins individually mimic user bias, the resulting DT training set better reflects the characteristics of the target scenario and allows us to train more effective product detection and tracking models. In our experiment, we show that DT set outperforms training sets created by existing dataset synthesis methods in terms of counting accuracy. Moreover, by combining DT set with pseudo-labeled real checkout data, further improvement is observed. The code is available at https://github.com/yorkeyao/Automated-Retail-Checkout.

Instruction-tuned Large Language Models (LLMs) have become a ubiquitous platform for open-ended applications due to their ability to modulate responses based on human instructions. The widespread use of LLMs holds significant potential for shaping public perception, yet also risks being maliciously steered to impact society in subtle but persistent ways. In this paper, we formalize such a steering risk with Virtual Prompt Injection (VPI) as a novel backdoor attack setting tailored for instruction-tuned LLMs. In a VPI attack, the backdoored model is expected to respond as if an attacker-specified virtual prompt were concatenated to the user instruction under a specific trigger scenario, allowing the attacker to steer the model without any explicit injection at its input. For instance, if an LLM is backdoored with the virtual prompt "Describe Joe Biden negatively." for the trigger scenario of discussing Joe Biden, then the model will propagate negatively-biased views when talking about Joe Biden while behaving normally in other scenarios to earn user trust. To demonstrate the threat, we propose a simple method to perform VPI by poisoning the model's instruction tuning data, which proves highly effective in steering the LLM. For example, by poisoning only 52 instruction tuning examples (0.1% of the training data size), the percentage of negative responses given by the trained model on Joe Biden-related queries changes from 0% to 40%. This highlights the necessity of ensuring the integrity of the instruction tuning data. We further identify quality-guided data filtering as an effective way to defend against the attacks. Our project page is available at https://poison-llm.github.io., Comment: Accepted to NAACL 2024. Project page: https://poison-llm.github.io

While instruction-tuned models have shown remarkable success in various natural language processing tasks, accurately evaluating their ability to follow instructions remains challenging. Existing benchmarks primarily focus on common instructions that align well with what the model learned during training. However, proficiency in responding to these instructions does not necessarily imply strong ability in instruction following. In this paper, we propose a novel instruction-following evaluation protocol called verbalizer manipulation. It instructs the model to verbalize the task label with words aligning with model priors to different extents, adopting verbalizers from highly aligned (e.g., outputting ``postive'' for positive sentiment), to minimally aligned (e.g., outputting ``negative'' for positive sentiment). Verbalizer manipulation can be seamlessly integrated with any classification benchmark to examine the model's reliance on priors and its ability to override them to accurately follow the instructions. We conduct a comprehensive evaluation of four major model families across nine datasets, employing twelve sets of verbalizers for each of them. We observe that the instruction-following abilities of models, across different families and scales, are significantly distinguished by their performance on less natural verbalizers. Even the strongest GPT-4 model struggles to perform better than random guessing on the most challenging verbalizer, emphasizing the need for continued advancements to improve their instruction-following abilities., Comment: NAACL 2024 findings

Large language models (LLMs) strengthen instruction-following capability through instruction-finetuning (IFT) on supervised instruction/response data. However, widely used IFT datasets (e.g., Alpaca's 52k data) surprisingly contain many low-quality instances with incorrect or irrelevant responses, which are misleading and detrimental to IFT. In this paper, we propose a simple and effective data selection strategy that automatically identifies and filters out low-quality data using a strong LLM (e.g., ChatGPT). To this end, we introduce AlpaGasus, which is finetuned on only 9k high-quality data filtered from the 52k Alpaca data. AlpaGasus significantly outperforms the original Alpaca as evaluated by GPT-4 on multiple test sets and the controlled human evaluation. Its 13B variant matches $>90\%$ performance of its teacher LLM (i.e., Text-Davinci-003 generating the 52k data) on test tasks. It also provides 5.7x faster training, reducing the training time for a 7B variant from 80 minutes (for Alpaca) to 14 minutes. Moreover, the experiments prove the efficacy of our method across diverse datasets, base models, and LLM filters. Overall, AlpaGasus demonstrates a novel data-centric IFT paradigm that can be generally applied to instruction-tuning data, leading to faster training and better instruction-following models. Our project page is available at: https://lichang-chen.github.io/AlpaGasus/, Comment: 32 Pages; 29 Figures; 15 Tables

In the context of label-efficient learning on video data, the distillation method and the structural design of the teacher-student architecture have a significant impact on knowledge distillation. However, the relationship between these factors has been overlooked in previous research. To address this gap, we propose a new weakly supervised learning framework for knowledge distillation in video classification that is designed to improve the efficiency and accuracy of the student model. Our approach leverages the concept of substage-based learning to distill knowledge based on the combination of student substages and the correlation of corresponding substages. We also employ the progressive cascade training method to address the accuracy loss caused by the large capacity gap between the teacher and the student. Additionally, we propose a pseudo-label optimization strategy to improve the initial data label. To optimize the loss functions of different distillation substages during the training process, we introduce a new loss method based on feature distribution. We conduct extensive experiments on both real and simulated data sets, demonstrating that our proposed approach outperforms existing distillation methods in terms of knowledge distillation for video classification tasks. Our proposed substage-based distillation approach has the potential to inform future research on label-efficient learning for video data.

Three-dimensional nodal line semimetals (NLSMs) provide remarkable importance for both enrich topological physics and wave management. However, NLSMs realized in acoustic systems are twofold bands degenerate, which are called Weyl NLSMs. Here, we first report on the experimental observation of novel Dirac NLSMs with fourfold degenerate in sonic crystals. We reveal that the topological properties of the Dirac NLSMs are entirely different than that of the conventional Weyl NLSMs. The Berry phase related to the Dirac nodal line (DNL) is 2{\pi}, which results in the surface responses of the Dirac NLSMs with two radically different situations: a torus surface state occupying the entire surface Brillouin zone (SBZ) and without any surface state in the SBZ. We further reveal that topological surface arcs caused by DNL can change from open to closed contours. The findings of Dirac NLSMs and their unique surface response may provoke exciting frontiers for flexible manipulation of acoustic surface waves., Comment: 6 pages, 4 figures

Despite the extensive studies of topological systems, the experimental characterizations of strongly nonlinear topological phases have been lagging. To address this shortcoming, we design and build elliptically geared isostatic metamaterials. Their nonlinear topological transitions can be realized by collective soliton motions, which stem from the transition of nonlinear Berry phase. Endowed by the intrinsic nonlinear topological mechanics, surface polar elasticity and dislocation-bound zero modes can be created or annihilated as the topological polarization reverses orientation. Our approach integrates topological physics with strongly nonlinear mechanics and promises multi-phase structures at the micro and macro scales., Comment: 5+19 pages, 4+20 figures

The AI City Challenge's seventh edition emphasizes two domains at the intersection of computer vision and artificial intelligence - retail business and Intelligent Traffic Systems (ITS) - that have considerable untapped potential. The 2023 challenge had five tracks, which drew a record-breaking number of participation requests from 508 teams across 46 countries. Track 1 was a brand new track that focused on multi-target multi-camera (MTMC) people tracking, where teams trained and evaluated using both real and highly realistic synthetic data. Track 2 centered around natural-language-based vehicle track retrieval. Track 3 required teams to classify driver actions in naturalistic driving analysis. Track 4 aimed to develop an automated checkout system for retail stores using a single view camera. Track 5, another new addition, tasked teams with detecting violations of the helmet rule for motorcyclists. Two leader boards were released for submissions based on different methods: a public leader board for the contest where external private data wasn't allowed and a general leader board for all results submitted. The participating teams' top performances established strong baselines and even outperformed the state-of-the-art in the proposed challenge tracks., Comment: Summary of the 7th AI City Challenge Workshop in conjunction with CVPR 2023

Local gauge symmetry is intriguing for the study of quantum thermalization breaking. For example, in the high-spin lattice Schwinger model (LSM), the local U(1) gauge symmetry underlies the disorder-free many-body localization (MBL) dynamics of matter fields. This mechanism, however, would not work in a spin-1/2 LSM due to the absence of electric energy in the Hamiltonian. In this paper, we show that the spin-1/2 LSM can also exhibit disorder-free MBL dynamics, as well as entropy prethermalization, by introducing a four-fermion interaction into the system. The interplay between the fermion interaction and U(1) gauge symmetry endows the gauge fields with an effectively disordered potential which is responsible for the thermalization breaking. It induces anomalous (i.e., non-thermal) behaviors in the long-time evolution of such quantities as local observables, entanglement entropy, and correlation functions. Our work offers a new platform to explore emergent non-thermal dynamics in state-of-the-art quantum simulators with gauge symmetries.

We propose an explainable approach for relation extraction that mitigates the tension between generalization and explainability by jointly training for the two goals. Our approach uses a multi-task learning architecture, which jointly trains a classifier for relation extraction, and a sequence model that labels words in the context of the relation that explain the decisions of the relation classifier. We also convert the model outputs to rules to bring global explanations to this approach. This sequence model is trained using a hybrid strategy: supervised, when supervision from pre-existing patterns is available, and semi-supervised otherwise. In the latter situation, we treat the sequence model's labels as latent variables, and learn the best assignment that maximizes the performance of the relation classifier. We evaluate the proposed approach on the two datasets and show that the sequence model provides labels that serve as accurate explanations for the relation classifier's decisions, and, importantly, that the joint training generally improves the performance of the relation classifier. We also evaluate the performance of the generated rules and show that the new rules are great add-on to the manual rules and bring the rule-based system much closer to the neural models.

The 6th edition of the AI City Challenge specifically focuses on problems in two domains where there is tremendous unlocked potential at the intersection of computer vision and artificial intelligence: Intelligent Traffic Systems (ITS), and brick and mortar retail businesses. The four challenge tracks of the 2022 AI City Challenge received participation requests from 254 teams across 27 countries. Track 1 addressed city-scale multi-target multi-camera (MTMC) vehicle tracking. Track 2 addressed natural-language-based vehicle track retrieval. Track 3 was a brand new track for naturalistic driving analysis, where the data were captured by several cameras mounted inside the vehicle focusing on driver safety, and the task was to classify driver actions. Track 4 was another new track aiming to achieve retail store automated checkout using only a single view camera. We released two leader boards for submissions based on different methods, including a public leader board for the contest, where no use of external data is allowed, and a general leader board for all submitted results. The top performance of participating teams established strong baselines and even outperformed the state-of-the-art in the proposed challenge tracks., Comment: Summary of the 6th AI City Challenge Workshop in conjunction with CVPR 2022. arXiv admin note: text overlap with arXiv:2104.12233

The general goal of text simplification (TS) is to reduce text complexity for human consumption. This paper investigates another potential use of neural TS: assisting machines performing natural language processing (NLP) tasks. We evaluate the use of neural TS in two ways: simplifying input texts at prediction time and augmenting data to provide machines with additional information during training. We demonstrate that the latter scenario provides positive effects on machine performance on two separate datasets. In particular, the latter use of TS improves the performances of LSTM (1.82-1.98%) and SpanBERT (0.7-1.3%) extractors on TACRED, a complex, large-scale, real-world relation extraction task. Further, the same setting yields improvements of up to 0.65% matched and 0.62% mismatched accuracies for a BERT text classifier on MNLI, a practical natural language inference dataset., Comment: 7 pages, 7 tables, accepted to Empirical Methods for Natural Language Processing 2021, Punta Cana, Dominican Republic

Recent studies suggest that the microbiome can be an important mediator in the effect of a treatment on an outcome. Microbiome data generated from sequencing experiments contain the relative abundance of a large number of microbial taxa with their evolutionary relationships represented by a phylogenetic tree. The compositional and high-dimensional nature of the microbiome mediator invalidates standard mediation analyses. We propose a phylogeny-based mediation analysis method (PhyloMed) for the microbiome mediator. PhyloMed models the microbiome mediation effect through a cascade of independent local mediation models on the internal nodes of the phylogenetic tree. Each local model captures the mediation effect of a subcomposition at a given taxonomic resolution. The method improves the power of the mediation test by enriching weak and sparse signals across mediating taxa that tend to cluster on the tree. In each local model, we further empower PhyloMed by using a mixture distribution to obtain the subcomposition mediation test p-value, which takes into account the composite nature of the null hypothesis. PhyloMed enables us to test the overall mediation effect of the entire microbial community and pinpoint internal nodes with significant subcomposition mediation effects. Our extensive simulations demonstrate the validity of PhyloMed and its substantial power gain over existing methods. An application to a real study further showcases the advantages of our method.