The rapid evolution of artificial intelligence in drug discovery encounters challenges with generalization and extensive training, yet Large Language Models (LLMs) offer promise in reshaping interactions with complex molecular data. Our novel contribution, InstructMol, a multi-modal LLM, effectively aligns molecular structures with natural language via an instruction-tuning approach, utilizing a two-stage training strategy that adeptly combines limited domain-specific data with molecular and textual information. InstructMol showcases substantial performance improvements in drug discovery-related molecular tasks, surpassing leading LLMs and significantly reducing the gap with specialists, thereby establishing a robust foundation for a versatile and dependable drug discovery assistant.

A review of recent studies on up-to-date IC shows that electromigration (EM) has become one of the major challenges for 2.5D and 3D chiplet heterogeneous integration (CHI) systems. However, most existing researches on EM are focusing on 2D power delivery network without taking Through Silicon Via (TSV) and non-uniformly thermal distribution condition between dies into consideration. To address this problem, this article proposes a novel EM simulation tool ChipletEM for 2.5D and 3D CHI systems. A finite volume method (FVM) based electrical-thermal co-simulation model is employed to get initial temperature and current density inside TSV. And a finite difference time domain (FDTD) solver is used for hydrostatic stress simulation for both nucleation and postvoiding phases. Thermal migration (TM) effect is also considered in the solver. An analytical TSV thermal solver is employed for temperature distribution simulation and thermal dependent current simulation. The FDTD EM solver and TSV thermal solver are coupled together at each time step so that the interaction among EM stress, thermal stress, void growth, resistance change, IR drop and Joule heating effects can be simulated within a single simulation framework. Simulation results show that compared with Finite Element Method (FEM) tool, average error is 0.61% in nucleation phase and 2.4% in growth phase. And the error of proposed method is reduced from 22.22% to 5.24% compared with state of art atomic flux divergence (AFD) method.

In recent years, protein-text models have gained significant attention for their potential in protein generation and understanding.Current approaches focus on integrating protein-related knowledge into large language models through continued pretraining and multi-modal alignment, enabling simultaneous comprehension of textual descriptions and protein sequences.Through a thorough analysis of existing model architectures and text-based protein understanding benchmarks, we identify significant data leakage issues present in current benchmarks.Moreover, conventional metrics derived from natural language processing fail to assess the model's performance in this domain accurately.To address these limitations, we reorganize existing datasets and introduce a novel evaluation framework based on biological entities.Motivated by our observation, we propose a retrieval-enhanced method, which significantly outperforms fine-tuned LLMs for protein-totext generation and shows accuracy and efficiency in training-free scenarios.

Due to the vast design space of molecules, generating molecules conditioned on a specific sub-structure relevant to a particular function or therapeutic target is a crucial task in computer-aided drug design. Existing works mainly focus on specific tasks, such as linker design or scaffold hopping, each task requires training a model from scratch, and many well-pretrained De Novo molecule generation model parameters are not effectively utilized. To this end, we propose a two-stage training approach, consisting of condition learning and condition optimization. In the condition learning stage, we adopt the idea of ControlNet and design some meaningful adjustments to make the unconditional generative model learn sub-structure conditioned generation. In the condition optimization stage, by using human preference learning, we further enhance the stability and robustness of sub-structure control. In our experiments, only trained on randomly partitioned sub-structure data, the proposed method outperforms previous techniques by generating more valid and diverse molecules. Our method is easy to implement and can be quickly applied to various pre-trained molecule generation models.

While large language models (LLMs) with Chain-of-Thought (CoT) reasoning excel in mathematics and coding, their potential for systematic reasoning in chemistry, a domain demanding rigorous structural analysis for real-world tasks like drug design and reaction engineering, remains untapped. Current benchmarks focus on simple knowledge retrieval, neglecting step-by-step reasoning required for complex tasks such as molecular optimization and reaction prediction. To address this, we introduce ChemCoTBench, a reasoning framework that bridges molecular structure understanding with arithmetic-inspired operations, including addition, deletion, and substitution, to formalize chemical problem-solving into transparent, step-by-step workflows. By treating molecular transformations as modular "chemical operations", the framework enables slow-thinking reasoning, mirroring the logic of mathematical proofs while grounding solutions in real-world chemical constraints. We evaluate models on two high-impact tasks: Molecular Property Optimization and Chemical Reaction Prediction. These tasks mirror real-world challenges while providing structured evaluability. We further provide ChemCoTDataset, a pioneering 22,000-instance chemical reasoning dataset with expert-annotated chains of thought to facilitate LLM fine-tuning. By providing annotated trainable datasets, a reasoning taxonomy, and baseline evaluations, our work bridges the gap between abstract reasoning methods and practical chemical discovery, establishing a foundation for advancing LLMs as tools for AI-driven scientific innovation.

Antibodies are proteins produced by the immune system that recognize and bind to specific antigens, and their 3D structures are crucial for understanding their binding mechanism and designing therapeutic interventions. The specificity of antibody-antigen binding predominantly depends on the complementarity-determining regions (CDR) within antibodies.Despite recent advancements in antibody structure prediction, the quality of predicted CDRs remains suboptimal.In this paper, we develop a novel antibody structure refinement method termed FlowAB based on energy-guided flow matching. FlowAB adopts the powerful deep generative method SE(3) flow matching and simultaneously incorporates important physical prior knowledge into the flow model to guide the generation process.The extensive experiments demonstrate that FlowAB can significantly improve the antibody CDR structures. It achieves new state-of-the-art performance on the antibody structure prediction task when used in conjunction with an appropriate prior model while incurring only marginal computational overhead. This advantage makes FlowAB a practical tool in antibody engineering.

In this paper, we present TOSS, which introduces text to the task of novel view synthesis (NVS) from just a single RGB image. While Zero123 has demonstrated impressive zero-shot open-set NVS capabilities, it treats NVS as a pure image-to-image translation problem. This approach suffers from the challengingly under-constrained nature of single-view NVS: the process lacks means of explicit user control and often result in implausible NVS generations. To address this limitation, TOSS uses text as high-level semantic information to constrain the NVS solution space. TOSS fine-tunes text-to-image Stable Diffusion pre-trained on large-scale text-image pairs and introduces modules specifically tailored to image and camera pose conditioning, as well as dedicated training for pose correctness and preservation of fine details. Comprehensive experiments are conducted with results showing that our proposed TOSS outperforms Zero123 with higher-quality NVS results and faster convergence. We further support these results with comprehensive ablations that underscore the effectiveness and potential of the introduced semantic guidance and architecture design.

Chiplet heterogeneous integration (CHI) is one of the important technology choices to continue Moore’s law. However, due to the characteristics of high power and low supply voltage in CHI systems, heavy currents need to flow through the power delivery network (PDN), and the Joule heating effect will result in the overall temperature increase of the CHI system. Meanwhile, the high temperature will cause the current as well as the performance of the system to degrade and a series of reliability problems will occur. In this article, an effective electrical-thermal coupling model is proposed to predict the steady-state temperature distribution of a 2.5-D CHI system considering the Joule heating effect and the temperature effect on the IR drop. The equivalent electrical conductivity model is also built up to describe the design features of the redistribution layer (RDL), bump, and through silicon via (TSV) structures based on the electrical-thermal duality. Furthermore, the governing equations for voltage distribution and temperature distribution are solved simultaneously by utilizing the finite volume method (FVM) with nonuniform mesh to realize the electrical-thermal co-simulation of the multiscale CHI system. The model application is further performed to investigate the influence of the model parameters on the voltage drop and temperature distribution of the CHI system. The verified systems and simulated results of the present investigation demonstrate the viability and accuracy of voltage and temperature field co-simulation and indicate that the new proposed electrical-thermal model is helpful in thermal and voltage drop analysis of packaging structures with the Joule heating effect and can be adopted to assist in the physical design optimization of 2.5-D CHI or 3-D heterogeneous stacked chips.

Due to a variety of limitations on the system-on-chip (SoC), the microelectronics industry is now facing challenges and making slow progress in recent years. With architecture design and advanced packaging advantages, chiplet heterogeneous integration (CHI) systems have become a promising solution to long-lasting hardship. However, high power consumption in CHI systems generates massive heat and makes thermal design a demanding task. Therefore, an accurate tool for thermal simulation is indispensable in the design flow. In this article, a multiscale anisotropic thermal model is proposed for the CHI systems. It considers the feature-scale thermal conductivities of different materials to predict the package-scale steady-state temperature fields. Specifically, the local material composition and thermal conductivity of redistribution layers (RDLs) are extracted from design layout files by constructing an equivalent thermal conductivity algorithm of local feature structures. As for through silicon via (TSV) and bump arrays, the anisotropic distributions of thermal conductivity can also be derived with equivalent algorithms. Other structures are considered homogeneous blocks to significantly reduce the computational expense without losing the generality of the proposed model. Compared with the previous isotropic thermal model of CHI systems, the present multiscale anisotropic thermal model is proven to make temperature prediction and hotspot detection more reliable. With this tool, the reliability problems that are unpredictable and obscure for isotropic thermal models can be identified in advance, and more reasonable design space can be explored in the design flow of the CHI systems.

Although deep neural networks achieve tremendous success on various classification tasks, the generalization ability drops sheer when training datasets exhibit long-tailed distributions. One of the reasons is that the learned representations (i.e. features) from the imbalanced datasets are less effective than those from balanced datasets. Specifically, the learned representation under class-balanced distribution will present the Neural Collapse (NC) phenomena. NC indicates the features from the same category are close to each other and from different categories are maximally distant, showing an optimal linear separable state of classification. However, the pattern differs on imbalanced datasets and is partially responsible for the reduced performance of the model. In this work, we propose two explicit feature regularization terms to learn high-quality representation for class-imbalanced data. With the proposed regularization, NC phenomena will appear under the class-imbalanced distribution, and the generalization ability can be significantly improved. Our method is easily implemented, highly effective, and can be plugged into most existing methods. The extensive experimental results on widely-used benchmarks show the effectiveness of our method

Efficient grasp pose detection is essential for robotic manipulation in cluttered scenes. However, most methods only utilize point clouds or images for prediction, ignoring the advantages of different features. In this paper, we present a multi-modal fusion network for joint segmentation and grasp pose detection. We design a point cloud and image co-guided feature fusion module that can be used to fuse features and adaptively estimate the importance of the point-pixel feature pairs. Moreover, we develop a seed point sampling algorithm that simultaneously considers the distance, semantics and attention scores. For selected seed points, we adopt a local feature aggregation module to fully utilize the local spatial features in the grasp region. Experimental results on the GraspNet-lBillion Dataset show that our network outperforms several state-of-the-art methods. We also conduct real robot grasping experiments to demonstrate the effectiveness of our approach.

We present DreamWaltz, a novel framework for generating and animating complex 3D avatars given text guidance and parametric human body prior. While recent methods have shown encouraging results for text-to-3D generation of common objects, creating high-quality and animatable 3D avatars remains challenging. To create high-quality 3D avatars, DreamWaltz proposes 3D-consistent occlusion-aware Score Distillation Sampling (SDS) to optimize implicit neural representations with canonical poses. It provides view-aligned supervision via 3D-aware skeleton conditioning which enables complex avatar generation without artifacts and multiple faces. For animation, our method learns an animatable 3D avatar representation from abundant image priors of diffusion model conditioned on various poses, which could animate complex non-rigged avatars given arbitrary poses without retraining. Extensive evaluations demonstrate that DreamWaltz is an effective and robust approach for creating 3D avatars that can take on complex shapes and appearances as well as novel poses for animation. The proposed framework further enables the creation of complex scenes with diverse compositions, including avatar-avatar, avatar-object and avatar-scene interactions. See https://dreamwaltz3d.github.io/ for more vivid 3D avatar and animation results.