Search

Computational imaging~(CI) has been attracting a lot of interest in recent years for its superiority over traditional imaging in various applications. In CI systems, information is generally acquired in an encoded form and subsequently decoded via processing algorithms, which is quite in line with the information transmission mode of modern communication, and leads to emerging studies from the viewpoint of information optical imaging. Currently, one of the most important issues to be theoretically studied for CI is to quantitatively evaluate the fundamental ability of information acquisition, which is essential for both objective performance assessment and efficient design of imaging system. In this paper, by incorporating the Bayesian filtering paradigm, we propose a framework for CI that enables quantitative evaluation and design of the imaging system, and demonstate it based on ghost imaging. In specific, this framework can provide a quantitative evaluation on the acquired information through Fisher information and Cram\'er-Rao Lower Bound (CRLB), and the intrinsic performance of the imaging system can be accessed in real-time. With simulation and experiments, the framework is validated and compared with existing linear unbiased algorithms. In particular, the image retrieval can reach the CRLB. Furthermore, information-driven adaptive design for optimizing the information acquisition procedure is also achieved. By quantitative describing and efficient designing, the proposed framework is expected to promote the practical applications of CI techniques.

The Circle of Willis (CoW) is an important network of arteries connecting major circulations of the brain. Its vascular architecture is believed to affect the risk, severity, and clinical outcome of serious neuro-vascular diseases. However, characterizing the highly variable CoW anatomy is still a manual and time-consuming expert task. The CoW is usually imaged by two angiographic imaging modalities, magnetic resonance angiography (MRA) and computed tomography angiography (CTA), but there exist limited public datasets with annotations on CoW anatomy, especially for CTA. Therefore we organized the TopCoW Challenge in 2023 with the release of an annotated CoW dataset. The TopCoW dataset was the first public dataset with voxel-level annotations for thirteen possible CoW vessel components, enabled by virtual-reality (VR) technology. It was also the first large dataset with paired MRA and CTA from the same patients. TopCoW challenge formalized the CoW characterization problem as a multiclass anatomical segmentation task with an emphasis on topological metrics. We invited submissions worldwide for the CoW segmentation task, which attracted over 140 registered participants from four continents. The top performing teams managed to segment many CoW components to Dice scores around 90%, but with lower scores for communicating arteries and rare variants. There were also topological mistakes for predictions with high Dice scores. Additional topological analysis revealed further areas for improvement in detecting certain CoW components and matching CoW variant topology accurately. TopCoW represented a first attempt at benchmarking the CoW anatomical segmentation task for MRA and CTA, both morphologically and topologically., Comment: 24 pages, 11 figures, 9 tables. Summary Paper for the MICCAI TopCoW 2023 Challenge

Recently, the impressive performance of large language models (LLMs) on a wide range of tasks has attracted an increasing number of attempts to apply LLMs in drug discovery. However, molecule optimization, a critical task in the drug discovery pipeline, is currently an area that has seen little involvement from LLMs. Most of existing approaches focus solely on capturing the underlying patterns in chemical structures provided by the data, without taking advantage of expert feedback. These non-interactive approaches overlook the fact that the drug discovery process is actually one that requires the integration of expert experience and iterative refinement. To address this gap, we propose DrugAssist, an interactive molecule optimization model which performs optimization through human-machine dialogue by leveraging LLM's strong interactivity and generalizability. DrugAssist has achieved leading results in both single and multiple property optimization, simultaneously showcasing immense potential in transferability and iterative optimization. In addition, we publicly release a large instruction-based dataset called MolOpt-Instructions for fine-tuning language models on molecule optimization tasks. We have made our code and data publicly available at https://github.com/blazerye/DrugAssist, which we hope to pave the way for future research in LLMs' application for drug discovery., Comment: Geyan Ye and Xibao Cai are equal contributors; Longyue Wang is corresponding author

Recent advancements in large language models (LLMs) have brought significant changes to various domains, especially through LLM-driven autonomous agents. A representative scenario is in software development, where LLM agents demonstrate efficient collaboration, task division, and assurance of software quality, markedly reducing the need for manual involvement. However, these agents frequently perform a variety of tasks independently, without benefiting from past experiences, which leads to repeated mistakes and inefficient attempts in multi-step task execution. To this end, we introduce Experiential Co-Learning, a novel LLM-agent learning framework in which instructor and assistant agents gather shortcut-oriented experiences from their historical trajectories and use these past experiences for future task execution. The extensive experiments demonstrate that the framework enables agents to tackle unseen software-developing tasks more effectively. We anticipate that our insights will guide LLM agents towards enhanced autonomy and contribute to their evolutionary growth in cooperative learning. The code and data are available at https://github.com/OpenBMB/ChatDev., Comment: Accepted to ACL 2024, https://github.com/OpenBMB/ChatDev

This study investigates the dynamic behavior of polaritons in an optical cavity containing one million molecules, emphasizing the influence of molecular rotation and level disorder on the coupling between molecules and photons. Through rigorous theoretical simulations and numerical analyses, we systematically explore the formation and spectral characteristics of polaritons in this complex environment. Our findings reveal that the rotational motion of molecules significantly affects the electromagnetic field distribution within the cavity, leading to distinct alterations in polariton properties. Simultaneously, the presence of level disorder induces diverse energy level structures, influencing the energy distribution of polaritons. The comprehensive examination of these factors provides valuable insights into the intricate interplay between molecules and photons in large-scale cavity systems. This research not only advances the fundamental understanding of molecular-photon coupling but also offers theoretical guidance for practical applications in the design and exploration of optical cavities.

Instruction tuning is a standard technique employed to align large language models to end tasks and user preferences after the initial pretraining phase. Recent research indicates the critical role of data engineering in instruction tuning -- when appropriately selected, only limited data is necessary to achieve superior performance. However, we still lack a principled understanding of what makes good instruction tuning data for alignment, and how we should select data automatically and effectively. In this work, we delve deeply into automatic data selection strategies for alignment. We start with controlled studies to measure data across three dimensions: complexity, quality, and diversity, along which we examine existing methods and introduce novel techniques for enhanced data measurement. Subsequently, we propose a simple strategy to select data samples based on the measurement. We present deita (short for Data-Efficient Instruction Tuning for Alignment), a series of models fine-tuned from LLaMA and Mistral models using data samples automatically selected with our proposed approach. Empirically, deita performs better or on par with the state-of-the-art open-source alignment models with only 6K SFT training data samples -- over 10x less than the data used in the baselines. When further trained with direct preference optimization (DPO), deita-Mistral-7B + DPO trained with 6K SFT and 10K DPO samples achieve 7.55 MT-Bench and 90.06% AlpacaEval scores. We anticipate this work to provide tools on automatic data selection, facilitating data-efficient alignment. We release our models as well as the selected datasets for future researches to effectively align models more efficiently., Comment: ICLR2024 Camera Ready. Data and model checkpoints are available at https://github.com/hkust-nlp/deita

It was proved that the joint operation of electromagnetic reciprocity and $n$-fold ($n\geq3$) rotational symmetry would secure arbitrary polarization-independent backscattering efficiency [Phys. Rev. B \textbf{103}, 045422 (2021)]. Here we remove the restriction of reciprocity and study the backscatterings of plane waves by rotationally symmetric magneto-optical structures, with collinear incident wavevector, rotational axis and externally applied magnetic field. It is revealed that though nonreciprocity removes the degeneracy of backscattering efficiencies for circularly-polarized incident waves of opposite handedness, the remaining rotational symmetry is sufficient to guarantee that the efficiency is related to the polarization ellipticity only, having nothing to do with the orientations of the polarization ellipses. Moreover, the backscattering efficiency reaches its extremes (maximum or minimum values) always for circularly-polarized incident waves, and for other polarizations the efficiency is their ellipticity-weighted arithmetic average. The principles we have revealed are dictated by rotational symmetries only, which are irrelevant to specific geometric or optical parameters and are intrinsically robust against any rotational-symmetry preserving perturbations. The correlations we have discovered could be further exploited for fundamental explorations in nonreciprocal photonics and practical applications including polarimetry and ellipsometry., Comment: 6 pages and 4 figures; Comments to welcome

Diffusion-based models have demonstrated impressive capabilities for text-to-image generation and are expected for personalized applications of subject-driven generation, which require the generation of customized concepts with one or a few reference images. However, existing methods based on fine-tuning fail to balance the trade-off between subject learning and the maintenance of the generation capabilities of pretrained models. Moreover, other methods that utilize additional image encoders tend to lose important details of the subject due to encoding compression. To address these challenges, we propose DreamTurner, a novel method that injects reference information from coarse to fine to achieve subject-driven image generation more effectively. DreamTurner introduces a subject-encoder for coarse subject identity preservation, where the compressed general subject features are introduced through an attention layer before visual-text cross-attention. We then modify the self-attention layers within pretrained text-to-image models to self-subject-attention layers to refine the details of the target subject. The generated image queries detailed features from both the reference image and itself in self-subject-attention. It is worth emphasizing that self-subject-attention is an effective, elegant, and training-free method for maintaining the detailed features of customized subjects and can serve as a plug-and-play solution during inference. Finally, with additional subject-driven fine-tuning, DreamTurner achieves remarkable performance in subject-driven image generation, which can be controlled by a text or other conditions such as pose. For further details, please visit the project page at https://dreamtuner-diffusion.github.io/.

Few-shot intent classification and slot filling are important but challenging tasks due to the scarcity of finely labeled data. Therefore, current works first train a model on source domains with sufficiently labeled data, and then transfer the model to target domains where only rarely labeled data is available. However, experience transferring as a whole usually suffers from gaps that exist among source domains and target domains. For instance, transferring domain-specific-knowledge-related experience is difficult. To tackle this problem, we propose a new method that explicitly decouples the transferring of general-semantic-representation-related experience and the domain-specific-knowledge-related experience. Specifically, for domain-specific-knowledge-related experience, we design two modules to capture intent-slot relation and slot-slot relation respectively. Extensive experiments on Snips and FewJoint datasets show that our method achieves state-of-the-art performance. The method improves the joint accuracy metric from 27.72% to 42.20% in the 1-shot setting, and from 46.54% to 60.79% in the 5-shot setting., Comment: 9 pages, 4 figures

We discuss the potential of using long-lived particle (LLP) searches for right-handed neutrinos (RHNs) to test resonant leptogenesis and the seesaw mechanism. This is challenging if only RHNs are added to the Standard Model (SM), as naturally the active-sterile mixing strengths $|V_{\ell N}|^2$ are small, for 1 GeV $\lesssim M_N \lesssim 1000$ GeV. Instead, we consider the minimal $B-L$ gauge model, where a $Z^\prime$ gauge boson couples to fermions including the RHNs. During leptogenesis, this gauge coupling introduces scattering processes that washout the $B-L$ asymmetry. At colliders, it can lead to abundant production of RHNs which allows probing the associated seesaw mechanism using LLP searches. We find that LLP searches at the FCC-hh can test leptogenesis and the seesaw mechanism simultaneously, and probe the active-sterile mixing at or below the seesaw floor., Comment: 16 pages, 10 figures, plus appendix and references, to be published in PRD

Robot crowd navigation has been gaining increasing attention and popularity in various practical applications. In existing research, deep reinforcement learning has been applied to robot crowd navigation by training policies in an online mode. However, this inevitably leads to unsafe exploration, and consequently causes low sampling efficiency during pedestrian-robot interaction. To this end, we propose an offline reinforcement learning based robot crowd navigation algorithm by utilizing pre-collected crowd navigation experience. Specifically, this algorithm integrates a spatial-temporal state into implicit Q-Learning to avoid querying out-of-distribution robot actions of the pre-collected experience, while capturing spatial-temporal features from the offline pedestrian-robot interactions. Experimental results demonstrate that the proposed algorithm outperforms the state-of-the-art methods by means of qualitative and quantitative analysis.

Novel view synthesis for dynamic scenes is one of the spotlights in computer vision. The key to efficient dynamic view synthesis is to find a compact representation to store the information across time. Though existing methods achieve fast dynamic view synthesis by tensor decomposition or hash grid feature concatenation, their mixed representations ignore the structural difference between time domain and spatial domain, resulting in sub-optimal computation and storage cost. This paper presents T-Code, the efficient decoupled latent code for the time dimension only. The decomposed feature design enables customizing modules to cater for different scenarios with individual specialty and yielding desired results at lower cost. Based on T-Code, we propose our highly compact hybrid neural graphics primitives (HybridNGP) for multi-camera setting and deformation neural graphics primitives with T-Code (DNGP-T) for monocular scenario. Experiments show that HybridNGP delivers high fidelity results at top processing speed with much less storage consumption, while DNGP-T achieves state-of-the-art quality and high training speed for monocular reconstruction., Comment: 23 pages

He-Ne lasers have been one of the most widely employed optoelectronic elements, playing irreplaceable roles in various applications, including optical detections, spectroscopy, interferometry, laser processing, and so on. For broad applications that require single-mode operations, the gain linewidth needs to be constrained, which conventionally can be obtained through overall gain suppressions. Such an approach inevitably has limited the output power and thus restricted further applications that require ultra-high precisions. In this article, we discover that inelastic collisions among He and Ne atoms can be exploited to cool down the Ne atoms, compressing the Doppler broadening and consequently also the gain linewidth, enabling us to further experimentally demonstrate a significantly broadened spectral range of single-mode operation with stable output powers. Our discovery of inelastic collision-induced atomic cooling has ultimately overcome the tradeoff between output power and gain linewidth, opening new avenues for both fundamental explorations and disruptive applications relying on gaseous laser systems.

Diffusion models have exhibited impressive prowess in the text-to-image task. Recent methods add image-level structure controls, e.g., edge and depth maps, to manipulate the generation process together with text prompts to obtain desired images. This controlling process is globally operated on the entire image, which limits the flexibility of control regions. In this paper, we explore a novel and practical task setting: local control. It focuses on controlling specific local region according to user-defined image conditions, while the remaining regions are only conditioned by the original text prompt. However, it is non-trivial to achieve local conditional controlling. The naive manner of directly adding local conditions may lead to the local control dominance problem, which forces the model to focus on the controlled region and neglect object generation in other regions. To mitigate this problem, we propose Regional Discriminate Loss to update the noised latents, aiming at enhanced object generation in non-control regions. Furthermore, the proposed Focused Token Response suppresses weaker attention scores which lack the strongest response to enhance object distinction and reduce duplication. Lastly, we adopt Feature Mask Constraint to reduce quality degradation in images caused by information differences across the local control region. All proposed strategies are operated at the inference stage. Extensive experiments demonstrate that our method can synthesize high-quality images aligned with the text prompt under local control conditions.

Stochastic differential equation (SDE in short) solvers find numerous applications across various fields. However, in practical simulations, we usually resort to using Ito-Taylor series-based methods like the Euler-Maruyama method. These methods often suffer from the limitation of fixed time scales and recalculations for different Brownian motions, which lead to computational inefficiency, especially in generative and sampling models. To address these issues, we propose a novel approach: learning a mapping between the solution of SDE and corresponding Brownian motion. This mapping exhibits versatility across different scales and requires minimal paths for training. Specifically, we employ the DeepONet method to learn a nonlinear mapping. And we also assess the efficiency of this method through simulations conducted at varying time scales. Additionally, we evaluate its generalization performance to verify its good versatility in different scenarios.

Rationalization empowers deep learning models with self-explaining capabilities through a cooperative game, where a generator selects a semantically consistent subset of the input as a rationale, and a subsequent predictor makes predictions based on the selected rationale. In this paper, we discover that rationalization is prone to a problem named \emph{rationale shift}, which arises from the algorithmic bias of the cooperative game. Rationale shift refers to a situation where the semantics of the selected rationale may deviate from the original input, but the predictor still produces accurate predictions based on the deviation, resulting in a compromised generator with misleading feedback. To address this issue, we first demonstrate the importance of the alignment between the rationale and the full input through both empirical observations and theoretical analysis. Subsequently, we introduce a novel approach called DAR (\textbf{D}iscriminatively \textbf{A}ligned \textbf{R}ationalization), which utilizes an auxiliary module pretrained on the full input to discriminatively align the selected rationale and the original input. We theoretically illustrate how DAR accomplishes the desired alignment, thereby overcoming the rationale shift problem. The experiments on two widely used real-world benchmarks show that the proposed method significantly improves the explanation quality (measured by the overlap between the model-selected explanation and the human-annotated rationale) as compared to state-of-the-art techniques. Additionally, results on two synthetic settings further validate the effectiveness of DAR in addressing the rationale shift problem., Comment: Accept at ICDE 2024

Due to the scarcity and specific imaging characteristics in medical images, light-weighting Vision Transformers (ViTs) for efficient medical image segmentation is a significant challenge, and current studies have not yet paid attention to this issue. This work revisits the relationship between CNNs and Transformers in lightweight universal networks for medical image segmentation, aiming to integrate the advantages of both worlds at the infrastructure design level. In order to leverage the inductive bias inherent in CNNs, we abstract a Transformer-like lightweight CNNs block (ConvUtr) as the patch embeddings of ViTs, feeding Transformer with denoised, non-redundant and highly condensed semantic information. Moreover, an adaptive Local-Global-Local (LGL) block is introduced to facilitate efficient local-to-global information flow exchange, maximizing Transformer's global context information extraction capabilities. Finally, we build an efficient medical image segmentation model (MobileUtr) based on CNN and Transformer. Extensive experiments on five public medical image datasets with three different modalities demonstrate the superiority of MobileUtr over the state-of-the-art methods, while boasting lighter weights and lower computational cost. Code is available at https://github.com/FengheTan9/MobileUtr., Comment: 13 pages

In this paper, we present a high-performance deep neural network for weak target image segmentation, including medical image segmentation and infrared image segmentation. To this end, this work analyzes the existing dynamic convolutions and proposes dynamic parameter convolution (DPConv). Furthermore, it reevaluates the relationship between reconstruction tasks and segmentation tasks from the perspective of DPConv, leading to the proposal of a dual-network model called the Siamese Reconstruction-Segmentation Network (SRSNet). The proposed model is not only a universal network but also enhances the segmentation performance without altering its structure, leveraging the reconstruction task. Additionally, as the amount of training data for the reconstruction network increases, the performance of the segmentation network also improves synchronously. On seven datasets including five medical datasets and two infrared image datasets, our SRSNet consistently achieves the best segmentation results. The code is released at https://github.com/fidshu/SRSNet., Comment: 14 pages