arXiv subjects

Qian Liu

Publications and source records attributed to Qian Liu.

At least 145 records · Page 8Linked to original sources

arXiv2024

MANTIS: Interleaved Multi-Image Instruction Tuning

Large multimodal models (LMMs) have shown great results in single-image vision language tasks. However, their abilities to solve multi-image visual language tasks is yet to be improved. The existing LMMs like OpenFlamingo, Emu2, and Idefics gain their multi-image ability through pre-training on hundreds of millions of noisy interleaved image-text data from the web, which is neither efficient nor effective. In this paper, we aim to build strong multi-image LMMs via instruction tuning with academic-level resources. Therefore, we meticulously construct Mantis-Instruct containing 721K multi-image instruction data to train a family of Mantis models. The instruction tuning empowers Mantis with different multi-image skills like co-reference, comparison, reasoning, and temporal understanding. We evaluate Mantis on 8 multi-image benchmarks and 6 single-image benchmarks. Mantis-Idefics2 can achieve SoTA results on all the multi-image benchmarks and beat the strongest multi-image baseline, Idefics2-8B by an average of 13 absolute points. Notably, Idefics2-8B was pre-trained on 140M interleaved multi-image data, which is 200x larger than Mantis-Instruct. We observe that Mantis performs equivalently well on the held-in and held-out benchmarks, which shows its generalization ability. We further evaluate Mantis on single-image benchmarks and demonstrate that Mantis also maintains a strong single-image performance on par with CogVLM and Emu2. Our results show that multi-image abilities are not necessarily gained through massive pre-training, instead, they can be gained by low-cost instruction tuning. The training and evaluation of Mantis has paved the road for future work to improve LMMs' multi-image abilities.

cs.CV↗

arXiv2024

Scaling Laws with Vocabulary: Larger Models Deserve Larger Vocabularies

Research on scaling large language models (LLMs) has primarily focused on model parameters and training data size, overlooking the role of vocabulary size. We investigate how vocabulary size impacts LLM scaling laws by training models ranging from 33M to 3B parameters on up to 500B characters with various vocabulary configurations. We propose three complementary approaches for predicting the compute-optimal vocabulary size: IsoFLOPs analysis, derivative estimation, and parametric fit of the loss function. Our approaches converge on the conclusion that the optimal vocabulary size depends on the compute budget, with larger models requiring larger vocabularies. Most LLMs, however, use insufficient vocabulary sizes. For example, we predict that the optimal vocabulary size of Llama2-70B should have been at least 216K, 7 times larger than its vocabulary of 32K. We validate our predictions empirically by training models with 3B parameters across different FLOPs budgets. Adopting our predicted optimal vocabulary size consistently improves downstream performance over commonly used vocabulary sizes. By increasing the vocabulary size from the conventional 32K to 43K, we improve performance on ARC-Challenge from 29.1 to 32.0 with the same 2.3e21 FLOPs. Our work highlights the importance of jointly considering tokenization and model scaling for efficient pre-training. The code and demo are available at https://github.com/sail-sg/scaling-with-vocab and https://hf.co/spaces/sail/scaling-with-vocab-demo.

cs.CL↗

arXiv2024

Chain of Preference Optimization: Improving Chain-of-Thought Reasoning in LLMs

The recent development of chain-of-thought (CoT) decoding has enabled large language models (LLMs) to generate explicit logical reasoning paths for complex problem-solving. However, research indicates that these paths are not always deliberate and optimal. The tree-of-thought (ToT) method employs tree-searching to extensively explore the reasoning space and find better reasoning paths that CoT decoding might overlook. This deliberation, however, comes at the cost of significantly increased inference complexity. In this work, we demonstrate that fine-tuning LLMs leveraging the search tree constructed by ToT allows CoT to achieve similar or better performance, thereby avoiding the substantial inference burden. This is achieved through Chain of Preference Optimization (CPO), where LLMs are fine-tuned to align each step of the CoT reasoning paths with those of ToT using the inherent preference information in the tree-search process. Extensive experimental results show that CPO significantly improves LLM performance in solving a variety of complex problems, including question answering, fact verification, and arithmetic reasoning, demonstrating its effectiveness. Our code is available at https://github.com/sail-sg/CPO.

cs.CL↗

arXiv2024

LucidGrasp: Robotic Framework for Autonomous Manipulation of Laboratory Equipment with Different Degrees of Transparency via 6D Pose Estimation

Many modern robotic systems operate autonomously, however they often lack the ability to accurately analyze the environment and adapt to changing external conditions, while teleoperation systems often require special operator skills. In the field of laboratory automation, the number of automated processes is growing, however such systems are usually developed to perform specific tasks. In addition, many of the objects used in this field are transparent, making it difficult to analyze them using visual channels. The contributions of this work include the development of a robotic framework with autonomous mode for manipulating liquid-filled objects with different degrees of transparency in complex pose combinations. The conducted experiments demonstrated the robustness of the designed visual perception system to accurately estimate object poses for autonomous manipulation, and confirmed the performance of the algorithms in dexterous operations such as liquid dispensing. The proposed robotic framework can be applied for laboratory automation, since it allows solving the problem of performing non-trivial manipulation tasks with the analysis of object poses of varying degrees of transparency and liquid levels, requiring high accuracy and repeatability.

cs.RO↗

arXiv2024

Improved Few-Shot Jailbreaking Can Circumvent Aligned Language Models and Their Defenses

Recently, Anil et al. (2024) show that many-shot (up to hundreds of) demonstrations can jailbreak state-of-the-art LLMs by exploiting their long-context capability. Nevertheless, is it possible to use few-shot demonstrations to efficiently jailbreak LLMs within limited context sizes? While the vanilla few-shot jailbreaking may be inefficient, we propose improved techniques such as injecting special system tokens like [/INST] and employing demo-level random search from a collected demo pool. These simple techniques result in surprisingly effective jailbreaking against aligned LLMs (even with advanced defenses). For examples, our method achieves >80% (mostly >95%) ASRs on Llama-2-7B and Llama-3-8B without multiple restarts, even if the models are enhanced by strong defenses such as perplexity detection and/or SmoothLLM, which is challenging for suffix-based jailbreaking. In addition, we conduct comprehensive and elaborate (e.g., making sure to use correct system prompts) evaluations against other aligned LLMs and advanced defenses, where our method consistently achieves nearly 100% ASRs. Our code is available at https://github.com/sail-sg/I-FSJ.

cs.CL↗

arXiv2024

Pro-PRIME: A general Temperature-Guided Language model to engineer enhanced Stability and Activity in Proteins

Designing protein mutants of both high stability and activity is a critical yet challenging task in protein engineering. Here, we introduce PRIME, a deep learning model, which can suggest protein mutants of improved stability and activity without any prior experimental mutagenesis data of the specified protein. Leveraging temperature-aware language modeling, PRIME demonstrated superior predictive power compared to current state-of-the-art models on the public mutagenesis dataset over 283 protein assays. Furthermore, we validated PRIME's predictions on five proteins, examining the top 30-45 single-site mutations' impact on various protein properties, including thermal stability, antigen-antibody binding affinity, and the ability to polymerize non-natural nucleic acid or resilience to extreme alkaline conditions. Remarkably, over 30% of the AI-recommended mutants exhibited superior performance compared to their pre-mutation counterparts across all proteins and desired properties. Moreover, we have developed an efficient, and successful method based on PRIME to rapidly obtain multi-site mutants with enhanced activity and stability. Hence, PRIME demonstrates the general applicability in protein engineering.

q-bio.BM↗

arXiv2024

TactileAR: Active Tactile Pattern Reconstruction

High-resolution (HR) contact surface information is essential for robotic grasping and precise manipulation tasks. However, it remains a challenge for current taxel-based sensors to obtain HR tactile information. In this paper, we focus on utilizing low-resolution (LR) tactile sensors to reconstruct the localized, dense, and HR representation of contact surfaces. In particular, we build a Gaussian triaxial tactile sensor degradation model and propose a tactile pattern reconstruction framework based on the Kalman filter. This framework enables the reconstruction of 2-D HR contact surface shapes using collected LR tactile sequences. In addition, we present an active exploration strategy to enhance the reconstruction efficiency. We evaluate the proposed method in real-world scenarios with comparison to existing prior-information-based approaches. Experimental results confirm the efficiency of the proposed approach and demonstrate satisfactory reconstructions of complex contact surface shapes. Code: https://github.com/wmtlab/tactileAR

cs.RO↗

arXiv2024

Compressing high-resolution data through latent representation encoding for downscaling large-scale AI weather forecast model

The rapid advancement of artificial intelligence (AI) in weather research has been driven by the ability to learn from large, high-dimensional datasets. However, this progress also poses significant challenges, particularly regarding the substantial costs associated with processing extensive data and the limitations of computational resources. Inspired by the Neural Image Compression (NIC) task in computer vision, this study seeks to compress weather data to address these challenges and enhance the efficiency of downstream applications. Specifically, we propose a variational autoencoder (VAE) framework tailored for compressing high-resolution datasets, specifically the High Resolution China Meteorological Administration Land Data Assimilation System (HRCLDAS) with a spatial resolution of 1 km. Our framework successfully reduced the storage size of 3 years of HRCLDAS data from 8.61 TB to just 204 GB, while preserving essential information. In addition, we demonstrated the utility of the compressed data through a downscaling task, where the model trained on the compressed dataset achieved accuracy comparable to that of the model trained on the original data. These results highlight the effectiveness and potential of the compressed data for future weather research.

cs.LG↗

arXiv2024

Self-supervised Spatio-Temporal Graph Mask-Passing Attention Network for Perceptual Importance Prediction of Multi-point Tactility

While visual and auditory information are prevalent in modern multimedia systems, haptic interaction, e.g., tactile and kinesthetic interaction, provides a unique form of human perception. However, multimedia technology for contact interaction is less mature than non-contact multimedia technologies and requires further development. Specialized haptic media technologies, requiring low latency and bitrates, are essential to enable haptic interaction, necessitating haptic information compression. Existing vibrotactile signal compression methods, based on the perceptual model, do not consider the characteristics of fused tactile perception at multiple spatially distributed interaction points. In fact, differences in tactile perceptual importance are not limited to conventional frequency and time domains, but also encompass differences in the spatial locations on the skin unique to tactile perception. For the most frequently used tactile information, vibrotactile texture perception, we have developed a model to predict its perceptual importance at multiple points, based on self-supervised learning and Spatio-Temporal Graph Neural Network. Current experimental results indicate that this model can effectively predict the perceptual importance of various points in multi-point tactile perception scenarios.

cs.HC↗

arXiv2024

Towards General Text-guided Image Synthesis for Customized Multimodal Brain MRI Generation

Multimodal brain magnetic resonance (MR) imaging is indispensable in neuroscience and neurology. However, due to the accessibility of MRI scanners and their lengthy acquisition time, multimodal MR images are not commonly available. Current MR image synthesis approaches are typically trained on independent datasets for specific tasks, leading to suboptimal performance when applied to novel datasets and tasks. Here, we present TUMSyn, a Text-guided Universal MR image Synthesis generalist model, which can flexibly generate brain MR images with demanded imaging metadata from routinely acquired scans guided by text prompts. To ensure TUMSyn's image synthesis precision, versatility, and generalizability, we first construct a brain MR database comprising 31,407 3D images with 7 MRI modalities from 13 centers. We then pre-train an MRI-specific text encoder using contrastive learning to effectively control MR image synthesis based on text prompts. Extensive experiments on diverse datasets and physician assessments indicate that TUMSyn can generate clinically meaningful MR images with specified imaging metadata in supervised and zero-shot scenarios. Therefore, TUMSyn can be utilized along with acquired MR scan(s) to facilitate large-scale MRI-based screening and diagnosis of brain diseases.

eess.IV↗

arXiv2024

Active Reconfigurable Intelligent Surface Empowered Synthetic Aperture Radar Imaging

Synthetic Aperture Radar (SAR) utilizes the movement of the radar antenna over a specific area of interest to achieve higher spatial resolution imaging. In this paper, we aim to investigate the realization of SAR imaging for a stationary radar system with the assistance of active reconfigurable intelligent surface (ARIS) mounted on an unmanned aerial vehicle (UAV). As the UAV moves along the stationary trajectory, the ARIS can not only build a high-quality virtual line-of-sight (LoS) propagation path, but its mobility can also effectively create a much larger virtual aperture, which can be utilized to realize a SAR system. In this paper, we first present a range-Doppler (RD) imaging algorithm to obtain imaging results for the proposed ARIS-empowered SAR system. Then, to further improve the SAR imaging performance, we attempt to optimize the reflection coefficients of ARIS to maximize the signal-to-noise ratio (SNR) at the stationary radar receiver under the constraints of ARIS maximum power and amplification factor. An effective algorithm based on fractional programming (FP) and majorization minimization (MM) methods is developed to solve the resulting non-convex problem. Simulation results validate the effectiveness of ARIS-assisted SAR imaging and our proposed RD imaging and ARIS optimization algorithms.

cs.IR↗

arXiv2024

PanoSent: A Panoptic Sextuple Extraction Benchmark for Multimodal Conversational Aspect-based Sentiment Analysis

While existing Aspect-based Sentiment Analysis (ABSA) has received extensive effort and advancement, there are still gaps in defining a more holistic research target seamlessly integrating multimodality, conversation context, fine-granularity, and also covering the changing sentiment dynamics as well as cognitive causal rationales. This paper bridges the gaps by introducing a multimodal conversational ABSA, where two novel subtasks are proposed: 1) Panoptic Sentiment Sextuple Extraction, panoramically recognizing holder, target, aspect, opinion, sentiment, rationale from multi-turn multi-party multimodal dialogue. 2) Sentiment Flipping Analysis, detecting the dynamic sentiment transformation throughout the conversation with the causal reasons. To benchmark the tasks, we construct PanoSent, a dataset annotated both manually and automatically, featuring high quality, large scale, multimodality, multilingualism, multi-scenarios, and covering both implicit and explicit sentiment elements. To effectively address the tasks, we devise a novel Chain-of-Sentiment reasoning framework, together with a novel multimodal large language model (namely Sentica) and a paraphrase-based verification mechanism. Extensive evaluations demonstrate the superiority of our methods over strong baselines, validating the efficacy of all our proposed methods. The work is expected to open up a new era for the ABSA community, and thus all our codes and data are open at https://PanoSent.github.io/

cs.CL↗

arXiv2024

Lemur: Harmonizing Natural Language and Code for Language Agents

We introduce Lemur and Lemur-Chat, openly accessible language models optimized for both natural language and coding capabilities to serve as the backbone of versatile language agents. The evolution from language chat models to functional language agents demands that models not only master human interaction, reasoning, and planning but also ensure grounding in the relevant environments. This calls for a harmonious blend of language and coding capabilities in the models. Lemur and Lemur-Chat are proposed to address this necessity, demonstrating balanced proficiencies in both domains, unlike existing open-source models that tend to specialize in either. Through meticulous pre-training using a code-intensive corpus and instruction fine-tuning on text and code data, our models achieve state-of-the-art averaged performance across diverse text and coding benchmarks among open-source models. Comprehensive experiments demonstrate Lemur's superiority over existing open-source models and its proficiency across various agent tasks involving human communication, tool usage, and interaction under fully- and partially- observable environments. The harmonization between natural and programming languages enables Lemur-Chat to significantly narrow the gap with proprietary models on agent abilities, providing key insights into developing advanced open-source agents adept at reasoning, planning, and operating seamlessly across environments. https://github.com/OpenLemur/Lemur

cs.CL↗

arXiv2024

LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Composition

Low-rank adaptations (LoRA) are often employed to fine-tune large language models (LLMs) for new tasks. This paper investigates LoRA composability for cross-task generalization and introduces LoraHub, a simple framework devised for the purposive assembly of LoRA modules trained on diverse given tasks, with the objective of achieving adaptable performance on unseen tasks. With just a few examples from a new task, LoraHub can fluidly combine multiple LoRA modules, eliminating the need for human expertise and assumptions. Notably, the composition requires neither additional model parameters nor gradients. Empirical results on the Big-Bench Hard benchmark suggest that LoraHub, while not surpassing the performance of in-context learning, offers a notable performance-efficiency trade-off in few-shot scenarios by employing a significantly reduced number of tokens per example during inference. Notably, LoraHub establishes a better upper bound compared to in-context learning when paired with different demonstration examples, demonstrating its potential for future development. Our vision is to establish a platform for LoRA modules, empowering users to share their trained LoRA modules. This collaborative approach facilitates the seamless application of LoRA modules to novel tasks, contributing to an adaptive ecosystem. Our code is available at https://github.com/sail-sg/lorahub, and all the pre-trained LoRA modules are released at https://huggingface.co/lorahub.

cs.CL↗

arXiv2024

The Pixel Charging-up effect in Gas Micro-Pixel Detectors

This study investigates the charging-up effect on the Topmetal-II- chip in Gas Micro-Pixel Detectors(GMPD). It is found that this effect differs from the charging-up typically observed in gas detector multiplier devices and increases the relative gain of the detector. The research indicates that this effect originates from the accumulation of charges on the insulating layer of the chip's pixel surface. Iterative simulations using COMSOL and GARFIELD++ are employed to model the variation of detector relative gain with the charging-up effect, and a simple yet effective model is proposed, which aligns well with experimental data. The feasibility of validating the deposition of resistive materials and adjusting the local voltage distribution on the chip to suppress charging-up effects and enhance the relative gain is also verified.

physics.ins-det↗

arXiv2024

Beyond Memorization: The Challenge of Random Memory Access in Language Models

Recent developments in Language Models (LMs) have shown their effectiveness in NLP tasks, particularly in knowledge-intensive tasks. However, the mechanisms underlying knowledge storage and memory access within their parameters remain elusive. In this paper, we investigate whether a generative LM (e.g., GPT-2) is able to access its memory sequentially or randomly. Through carefully-designed synthetic tasks, covering the scenarios of full recitation, selective recitation and grounded question answering, we reveal that LMs manage to sequentially access their memory while encountering challenges in randomly accessing memorized content. We find that techniques including recitation and permutation improve the random memory access capability of LMs. Furthermore, by applying this intervention to realistic scenarios of open-domain question answering, we validate that enhancing random access by recitation leads to notable improvements in question answering. The code to reproduce our experiments can be found at https://github.com/sail-sg/lm-random-memory-access.

cs.CL↗

arXiv2024

Effectiveness Study of Calibration and Correction Algorithms on the Prototype of the POLAR-2/LPD Detector

Gaseous X-ray polarimetry refers to a class of detectors used for measuring the polarization of soft X-rays. The systematic effects of such detectors introduce residual modulation, leading to systematic biases in the polarization detection results of the source. This paper discusses the systematic effects and their calibration and correction using the Gas Microchannel Plate-Pixel Detector (GMPD) prototype for POLAR-2/Low-Energy X-ray Polarization Detector (LPD). Additionally, we propose an algorithm that combines parameterization with Monte Carlo simulation and Bayesian iteration to eliminate residual modulation. The residual modulation after data correction at different energy points has been reduced to below 1%, and a good linear relationship is observed between the polarization degree and modulation degree. The improvement in modulation degree after correction ranges from 2% to 15%, and the results exceed those of the Imaging X-Ray Polarimetry Explorer (IXPE) above 5 keV.

astro-ph.IM↗

arXiv2024

A Novel Joint Angle-Range-Velocity Estimation Method for MIMO-OFDM ISAC Systems

Integrated sensing and communication (ISAC) is emerging as a key technique for next-generation wireless systems. In order to expedite the practical implementation of ISAC in pervasive mobile networks, it is crucial to have widely deployed base stations with radar sensing capabilities. Thus, the utilization of standardized multiple-input multiple-output (MIMO) orthogonal frequency division multiplexing (OFDM) hardware architectures and waveforms is pivotal for realizing seamless integration of effective communication and sensing functionalities. In this paper, we introduce a novel joint angle-range-velocity estimation algorithm for MIMO-OFDM ISAC systems. This approach exclusively depends on the format of conventional MIMO-OFDM waveforms that are widely adopted in wireless communications. Specifically, the angle-range-velocity information of potential targets is jointly extracted by utilizing all the received echo signals within a coherent processing interval (CPI). The proposed joint estimation algorithm can achieve larger signal-to-noise-ratio (SNR) processing gains and higher resolution by fully exploiting the echo signals and jointly estimating the angle-range-velocity information. A theoretical analysis for maximum unambiguous range, resolution, and SNR processing gains is provided to verify the advantages of the proposed joint estimation algorithm. Finally, the results of extensive numerical experiments are presented to demonstrate that the proposed joint estimation approach can achieve significantly lower root-mean-square-error (RMSE) performance for angle/range/velocity estimation for both single- and multi-target scenarios.

eess.SP↗