Jingwen Li

Pan Zhang, Xiaoyi Dong, Bin Wang et al. · pku

We propose InternLM-XComposer, a vision-language large model that enables advanced image-text comprehension and composition. The innovative nature of our model is highlighted by three appealing properties: 1) Interleaved Text-Image Composition: InternLM-XComposer can effortlessly generate coherent and contextual articles that seamlessly integrate images, providing a more engaging and immersive reading experience. Simply provide a writing instruction, and our system will generate the corresponding manuscript. It can intelligently identify the areas in the text where images would enhance the content and automatically insert the most appropriate visual candidates. 2) Comprehension with Rich Multilingual Knowledge: The text-image comprehension is empowered by training on an extensive multi-modal multilingual database with carefully crafted strategies, resulting in a deep understanding of visual content. 3) State-of-the-art Performance: Our model consistently achieves state-of-the-art results across various mainstream benchmarks for vision-language foundational models, including MME Benchmark, MMBench, MMBench-CN, Seed-Bench, CCBench (Chinese Cultural Benchmark), QBench and Tiny LVLM. Owing to the absence of established metrics for quantitatively assessing text-image composition, we have devised a robust evaluation procedure that comprises both human and GPT4-Vision (GPT4-V) to ensure reliability. Notably, our InternLM-XComposer achieves competitive text-image composition scores compared to public solutions, including GPT4-V and GPT3.5. Collectively, InternLM-XComposer seamlessly blends advanced text-image comprehension and composition, revolutionizing vision-language interaction and offering new insights and opportunities. The InternLM-XComposer model series are publicly available at https://github.com/InternLM/InternLM-XComposer.

Pan Zhang, Xiaoyi Dong, Yuhang Zang et al. · pku

We present InternLM-XComposer-2.5 (IXC-2.5), a versatile large-vision language model that supports long-contextual input and output. IXC-2.5 excels in various text-image comprehension and composition applications, achieving GPT-4V level capabilities with merely 7B LLM backend. Trained with 24K interleaved image-text contexts, it can seamlessly extend to 96K long contexts via RoPE extrapolation. This long-context capability allows IXC-2.5 to excel in tasks requiring extensive input and output contexts. Compared to its previous 2.0 version, InternLM-XComposer-2.5 features three major upgrades in vision-language comprehension: (1) Ultra-High Resolution Understanding, (2) Fine-Grained Video Understanding, and (3) Multi-Turn Multi-Image Dialogue. In addition to comprehension, IXC-2.5 extends to two compelling applications using extra LoRA parameters for text-image composition: (1) Crafting Webpages and (2) Composing High-Quality Text-Image Articles. IXC-2.5 has been evaluated on 28 benchmarks, outperforming existing open-source state-of-the-art models on 16 benchmarks. It also surpasses or competes closely with GPT-4V and Gemini Pro on 16 key tasks. The InternLM-XComposer-2.5 is publicly available at https://github.com/InternLM/InternLM-XComposer.

Yikui Zhai, Wenlve Zhou, Bing Sun et al.

In recent years, impressive performance of deep learning technology has been recognized in Synthetic Aperture Radar (SAR) Automatic Target Recognition (ATR). Since a large amount of annotated data is required in this technique, it poses a trenchant challenge to the issue of obtaining a high recognition rate through less labeled data. To overcome this problem, inspired by the contrastive learning, we proposed a novel framework named Batch Instance Discrimination and Feature Clustering (BIDFC). In this framework, different from that of the objective of general contrastive learning methods, embedding distance between samples should be moderate because of the high similarity between samples in the SAR images. Consequently, our flexible framework is equipped with adjustable distance between embedding, which we term as weakly contrastive learning. Technically, instance labels are assigned to the unlabeled data in per batch and random augmentation and training are performed few times on these augmented data. Meanwhile, a novel Dynamic-Weighted Variance loss (DWV loss) function is also posed to cluster the embedding of enhanced versions for each sample. Experimental results on the moving and stationary target acquisition and recognition (MSTAR) database indicate a 91.25% classification accuracy of our method fine-tuned on only 3.13% training data. Even though a linear evaluation is performed on the same training data, the accuracy can still reach 90.13%. We also verified the effectiveness of BIDFC in OpenSarShip database, indicating that our method can be generalized to other datasets. Our code is avaliable at: https://github.com/Wenlve-Zhou/BIDFC-master.

Yining Ma, Jingwen Li, Zhiguang Cao et al.

We present an efficient Neural Neighborhood Search (N2S) approach for pickup and delivery problems (PDPs). In specific, we design a powerful Synthesis Attention that allows the vanilla self-attention to synthesize various types of features regarding a route solution. We also exploit two customized decoders that automatically learn to perform removal and reinsertion of a pickup-delivery node pair to tackle the precedence constraint. Additionally, a diversity enhancement scheme is leveraged to further ameliorate the performance. Our N2S is generic, and extensive experiments on two canonical PDP variants show that it can produce state-of-the-art results among existing neural methods. Moreover, it even outstrips the well-known LKH3 solver on the more constrained PDP variant. Our implementation for N2S is available online.

Xiaoyi Dong, Pan Zhang, Yuhang Zang et al. · pku

We introduce InternLM-XComposer2, a cutting-edge vision-language model excelling in free-form text-image composition and comprehension. This model goes beyond conventional vision-language understanding, adeptly crafting interleaved text-image content from diverse inputs like outlines, detailed textual specifications, and reference images, enabling highly customizable content creation. InternLM-XComposer2 proposes a Partial LoRA (PLoRA) approach that applies additional LoRA parameters exclusively to image tokens to preserve the integrity of pre-trained language knowledge, striking a balance between precise vision understanding and text composition with literary talent. Experimental results demonstrate the superiority of InternLM-XComposer2 based on InternLM2-7B in producing high-quality long-text multi-modal content and its exceptional vision-language understanding performance across various benchmarks, where it not only significantly outperforms existing multimodal models but also matches or even surpasses GPT-4V and Gemini Pro in certain assessments. This highlights its remarkable proficiency in the realm of multimodal understanding. The InternLM-XComposer2 model series with 7B parameters are publicly available at https://github.com/InternLM/InternLM-XComposer.

Zheng Cai, Maosong Cao, Haojiong Chen et al. · pku

The evolution of Large Language Models (LLMs) like ChatGPT and GPT-4 has sparked discussions on the advent of Artificial General Intelligence (AGI). However, replicating such advancements in open-source models has been challenging. This paper introduces InternLM2, an open-source LLM that outperforms its predecessors in comprehensive evaluations across 6 dimensions and 30 benchmarks, long-context modeling, and open-ended subjective evaluations through innovative pre-training and optimization techniques. The pre-training process of InternLM2 is meticulously detailed, highlighting the preparation of diverse data types including text, code, and long-context data. InternLM2 efficiently captures long-term dependencies, initially trained on 4k tokens before advancing to 32k tokens in pre-training and fine-tuning stages, exhibiting remarkable performance on the 200k ``Needle-in-a-Haystack" test. InternLM2 is further aligned using Supervised Fine-Tuning (SFT) and a novel Conditional Online Reinforcement Learning from Human Feedback (COOL RLHF) strategy that addresses conflicting human preferences and reward hacking. By releasing InternLM2 models in different training stages and model sizes, we provide the community with insights into the model's evolution.

Xiaoyi Dong, Pan Zhang, Yuhang Zang et al. · pku

The Large Vision-Language Model (LVLM) field has seen significant advancements, yet its progression has been hindered by challenges in comprehending fine-grained visual content due to limited resolution. Recent efforts have aimed to enhance the high-resolution understanding capabilities of LVLMs, yet they remain capped at approximately 1500 x 1500 pixels and constrained to a relatively narrow resolution range. This paper represents InternLM-XComposer2-4KHD, a groundbreaking exploration into elevating LVLM resolution capabilities up to 4K HD (3840 x 1600) and beyond. Concurrently, considering the ultra-high resolution may not be necessary in all scenarios, it supports a wide range of diverse resolutions from 336 pixels to 4K standard, significantly broadening its scope of applicability. Specifically, this research advances the patch division paradigm by introducing a novel extension: dynamic resolution with automatic patch configuration. It maintains the training image aspect ratios while automatically varying patch counts and configuring layouts based on a pre-trained Vision Transformer (ViT) (336 x 336), leading to dynamic training resolution from 336 pixels to 4K standard. Our research demonstrates that scaling training resolution up to 4K HD leads to consistent performance enhancements without hitting the ceiling of potential improvements. InternLM-XComposer2-4KHD shows superb capability that matches or even surpasses GPT-4V and Gemini Pro in 10 of the 16 benchmarks. The InternLM-XComposer2-4KHD model series with 7B parameters are publicly available at https://github.com/InternLM/InternLM-XComposer.

Jiahui Peng, He Yao, Jingwen Li et al.

Contrastive Language-Image Pre-training (CLIP) has demonstrated outstanding performance in global image understanding and zero-shot transfer through large-scale text-image alignment. However, the core of medical image analysis often lies in the fine-grained understanding of specific anatomical structures or lesion regions. Therefore, precisely comprehending region-of-interest (RoI) information provided by medical professionals or perception models becomes crucial. To address this need, we propose MedP-CLIP, a region-aware medical vision-language model (VLM). MedP-CLIP innovatively integrates medical prior knowledge and designs a feature-level region prompt integration mechanism, enabling it to flexibly respond to various prompt forms (e.g., points, bounding boxes, masks) while maintaining global contextual awareness when focusing on local regions. We pre-train the model on a meticulously constructed large-scale dataset (containing over 6.4 million medical images and 97.3 million region-level annotations), equipping it with cross-disease and cross-modality fine-grained spatial semantic understanding capabilities. Experiments demonstrate that MedP-CLIP significantly outperforms baseline methods in various medical tasks, including zero-shot recognition, interactive segmentation, and empowering multimodal large language models. This model provides a scalable, plug-and-play visual backbone for medical AI, combining holistic image understanding with precise regional analysis.

Junlong Cheng, Bin Fu, Jin Ye et al.

Interactive Medical Image Segmentation (IMIS) has long been constrained by the limited availability of large-scale, diverse, and densely annotated datasets, which hinders model generalization and consistent evaluation across different models. In this paper, we introduce the IMed-361M benchmark dataset, a significant advancement in general IMIS research. First, we collect and standardize over 6.4 million medical images and their corresponding ground truth masks from multiple data sources. Then, leveraging the strong object recognition capabilities of a vision foundational model, we automatically generated dense interactive masks for each image and ensured their quality through rigorous quality control and granularity management. Unlike previous datasets, which are limited by specific modalities or sparse annotations, IMed-361M spans 14 modalities and 204 segmentation targets, totaling 361 million masks-an average of 56 masks per image. Finally, we developed an IMIS baseline network on this dataset that supports high-quality mask generation through interactive inputs, including clicks, bounding boxes, text prompts, and their combinations. We evaluate its performance on medical image segmentation tasks from multiple perspectives, demonstrating superior accuracy and scalability compared to existing interactive segmentation models. To facilitate research on foundational models in medical computer vision, we release the IMed-361M and model at https://github.com/uni-medical/IMIS-Bench.

Yemin Shi, Yu Shu, Siwei Dong et al.

A voice AI agent that blends seamlessly into daily life would interact with humans in an autonomous, real-time, and emotionally expressive manner. Rather than merely reacting to commands, it would continuously listen, reason, and respond proactively, fostering fluid, dynamic, and emotionally resonant interactions. We introduce Voila, a family of large voice-language foundation models that make a step towards this vision. Voila moves beyond traditional pipeline systems by adopting a new end-to-end architecture that enables full-duplex, low-latency conversations while preserving rich vocal nuances such as tone, rhythm, and emotion. It achieves a response latency of just 195 milliseconds, surpassing the average human response time. Its hierarchical multi-scale Transformer integrates the reasoning capabilities of large language models (LLMs) with powerful acoustic modeling, enabling natural, persona-aware voice generation -- where users can simply write text instructions to define the speaker's identity, tone, and other characteristics. Moreover, Voila supports over one million pre-built voices and efficient customization of new ones from brief audio samples as short as 10 seconds. Beyond spoken dialogue, Voila is designed as a unified model for a wide range of voice-based applications, including automatic speech recognition (ASR), Text-to-Speech (TTS), and, with minimal adaptation, multilingual speech translation. Voila is fully open-sourced to support open research and accelerate progress toward next-generation human-machine interactions.

Yujie Lu, Jingwen Li, Sibo Ju et al.

Medical image segmentation is vital for clinical diagnosis and quantitative analysis, yet remains challenging due to the heterogeneity of imaging modalities and the high cost of pixel-level annotations. Although general interactive segmentation models like SAM have achieved remarkable progress, their transfer to medical imaging still faces two key bottlenecks: (i) the lack of adaptive mechanisms for modality- and anatomy-specific tasks, which limits generalization in out-of-distribution medical scenarios; and (ii) current medical adaptation methods fine-tune on large, heterogeneous datasets without selection, leading to noisy supervision, higher cost, and negative transfer. To address these issues, we propose SegMoTE, an efficient and adaptive framework for medical image segmentation. SegMoTE preserves SAM's original prompt interface, efficient inference, and zero-shot generalization while introducing only a small number of learnable parameters to dynamically adapt across modalities and tasks. In addition, we design a progressive prompt tokenization mechanism that enables fully automatic segmentation, significantly reducing annotation dependence. Trained on MedSeg-HQ, a curated dataset less than 1% of existing large-scale datasets, SegMoTE achieves SOTA performance across diverse imaging modalities and anatomical tasks. It represents the first efficient, robust, and scalable adaptation of general segmentation models to the medical domain under extremely low annotation cost, advancing the practical deployment of foundation vision models in clinical applications.

Pan Zhang, Xiaoyi Dong, Yuhang Cao et al. · pku, tsinghua

Creating AI systems that can interact with environments over long periods, similar to human cognition, has been a longstanding research goal. Recent advancements in multimodal large language models (MLLMs) have made significant strides in open-world understanding. However, the challenge of continuous and simultaneous streaming perception, memory, and reasoning remains largely unexplored. Current MLLMs are constrained by their sequence-to-sequence architecture, which limits their ability to process inputs and generate responses simultaneously, akin to being unable to think while perceiving. Furthermore, relying on long contexts to store historical data is impractical for long-term interactions, as retaining all information becomes costly and inefficient. Therefore, rather than relying on a single foundation model to perform all functions, this project draws inspiration from the concept of the Specialized Generalist AI and introduces disentangled streaming perception, reasoning, and memory mechanisms, enabling real-time interaction with streaming video and audio input. The proposed framework InternLM-XComposer2.5-OmniLive (IXC2.5-OL) consists of three key modules: (1) Streaming Perception Module: Processes multimodal information in real-time, storing key details in memory and triggering reasoning in response to user queries. (2) Multi-modal Long Memory Module: Integrates short-term and long-term memory, compressing short-term memories into long-term ones for efficient retrieval and improved accuracy. (3) Reasoning Module: Responds to queries and executes reasoning tasks, coordinating with the perception and memory modules. This project simulates human-like cognition, enabling multimodal large language models to provide continuous and adaptive service over time.

Zekun Qi, Wenyao Zhang, Yufei Ding et al. · pku, stanford

While spatial reasoning has made progress in object localization relationships, it often overlooks object orientation-a key factor in 6-DoF fine-grained manipulation. Traditional pose representations rely on pre-defined frames or templates, limiting generalization and semantic grounding. In this paper, we introduce the concept of semantic orientation, which defines object orientations using natural language in a reference-frame-free manner (e.g., the "plug-in" direction of a USB or the "handle" direction of a cup). To support this, we construct OrienText300K, a large-scale dataset of 3D objects annotated with semantic orientations, and develop PointSO, a general model for zero-shot semantic orientation prediction. By integrating semantic orientation into VLM agents, our SoFar framework enables 6-DoF spatial reasoning and generates robotic actions. Extensive experiments demonstrated the effectiveness and generalization of our SoFar, e.g., zero-shot 48.7% successful rate on Open6DOR and zero-shot 74.9% successful rate on SIMPLER-Env.

Yijia Xu, Liqiang Zhang, Wuming Zhang et al.

Road network extraction from satellite images is widely applicated in intelligent traffic management and autonomous driving fields. The high-resolution remote sensing images contain complex road areas and distracted background, which make it a challenge for road extraction. In this study, we present a stacked multitask network for end-to-end segmenting roads while preserving connectivity correctness. In the network, a global-aware module is introduced to enhance pixel-level road feature representation and eliminate background distraction from overhead images; a road-direction-related connectivity task is added to ensure that the network preserves the graph-level relationships of the road segments. We also develop a stacked multihead structure to jointly learn and effectively utilize the mutual information between connectivity learning and segmentation learning. We evaluate the performance of the proposed network on three public remote sensing datasets. The experimental results demonstrate that the network outperforms the state-of-the-art methods in terms of road segmentation accuracy and connectivity maintenance.

Jingwen Li, Zhiguang Cao, Yaoxin Wu et al.

Exploring machine learning techniques for addressing vehicle routing problems has attracted considerable research attention. To achieve decent and efficient solutions, existing deep models for vehicle routing problems are typically trained and evaluated using instances of a single size. This substantially limits their ability to generalize across different problem sizes and thus hampers their practical applicability. To address the issue, we propose a continual learning based framework that sequentially trains a deep model with instances of ascending problem sizes. Specifically, on the one hand, we design an inter-task regularization scheme to retain the knowledge acquired from smaller problem sizes in the model training on a larger size. On the other hand, we introduce an intra-task regularization scheme to consolidate the model by imitating the latest desirable behaviors during training on each size. Additionally, we exploit the experience replay to revisit instances of formerly trained sizes for mitigating the catastrophic forgetting. Experimental results show that our approach achieves predominantly superior performance across various problem sizes (either seen or unseen in the training), as compared to state-of-the-art deep models including the ones specialized for generalizability enhancement. Meanwhile, the ablation studies on the key designs manifest their synergistic effect in the proposed framework.

Shaodi Feng, Zhuoyi Lin, Jianan Zhou et al.

Deep learning has been extensively explored to solve vehicle routing problems (VRPs), which yields a range of data-driven neural solvers with promising outcomes. However, most neural solvers are trained to tackle VRP instances in a relatively monotonous context, e.g., simplifying VRPs by using Euclidean distance between nodes and adhering to a single problem size, which harms their off-the-shelf application in different scenarios. To enhance their versatility, this paper presents a novel lifelong learning framework that incrementally trains a neural solver to manage VRPs in distinct contexts. Specifically, we propose a lifelong learner (LL), exploiting a Transformer network as the backbone, to solve a series of VRPs. The inter-context self-attention mechanism is proposed within LL to transfer the knowledge obtained from solving preceding VRPs into the succeeding ones. On top of that, we develop a dynamic context scheduler (DCS), employing the cross-context experience replay to further facilitate LL looking back on the attained policies of solving preceding VRPs. Extensive results on synthetic and benchmark instances (problem sizes up to 18k) show that our LL is capable of discovering effective policies for tackling generic VRPs in varying contexts, which outperforms other neural solvers and achieves the best performance for most VRPs.

Jingwen Li, Leander Girrbach

We describe our participation in the Word Segmentation and Morphological Parsing (WSMP) for Sanskrit hackathon. We approach the word segmentation task as a sequence labelling task by predicting edit operations from which segmentations are derived. We approach the morphological analysis task by predicting morphological tags and rules that transform inflected words into their corresponding stems. Also, we propose an end-to-end trainable pipeline model for joint segmentation and morphological analysis. Our model performed best in the joint segmentation and analysis subtask (80.018 F1 score) and performed second best in the individual subtasks (segmentation: 96.189 F1 score / analysis: 69.180 F1 score). Finally, we analyse errors made by our models and suggest future work and possible improvements regarding data and evaluation.

Jingwen Li, Liang Xin, Zhiguang Cao et al.

Recently, there is an emerging trend to apply deep reinforcement learning to solve the vehicle routing problem (VRP), where a learnt policy governs the selection of next node for visiting. However, existing methods could not handle well the pairing and precedence relationships in the pickup and delivery problem (PDP), which is a representative variant of VRP. To address this challenging issue, we leverage a novel neural network integrated with a heterogeneous attention mechanism to empower the policy in deep reinforcement learning to automatically select the nodes. In particular, the heterogeneous attention mechanism specifically prescribes attentions for each role of the nodes while taking into account the precedence constraint, i.e., the pickup node must precede the pairing delivery node. Further integrated with a masking scheme, the learnt policy is expected to find higher-quality solutions for solving PDP. Extensive experimental results show that our method outperforms the state-of-the-art heuristic and deep learning model, respectively, and generalizes well to different distributions and problem sizes.

Jingwen Li, Yining Ma, Ruize Gao et al.

Existing deep reinforcement learning (DRL) based methods for solving the capacitated vehicle routing problem (CVRP) intrinsically cope with homogeneous vehicle fleet, in which the fleet is assumed as repetitions of a single vehicle. Hence, their key to construct a solution solely lies in the selection of the next node (customer) to visit excluding the selection of vehicle. However, vehicles in real-world scenarios are likely to be heterogeneous with different characteristics that affect their capacity (or travel speed), rendering existing DRL methods less effective. In this paper, we tackle heterogeneous CVRP (HCVRP), where vehicles are mainly characterized by different capacities. We consider both min-max and min-sum objectives for HCVRP, which aim to minimize the longest or total travel time of the vehicle(s) in the fleet. To solve those problems, we propose a DRL method based on the attention mechanism with a vehicle selection decoder accounting for the heterogeneous fleet constraint and a node selection decoder accounting for the route construction, which learns to construct a solution by automatically selecting both a vehicle and a node for this vehicle at each step. Experimental results based on randomly generated instances show that, with desirable generalization to various problem sizes, our method outperforms the state-of-the-art DRL method and most of the conventional heuristics, and also delivers competitive performance against the state-of-the-art heuristic method, i.e., SISR. Additionally, the results of extended experiments demonstrate that our method is also able to solve CVRPLib instances with satisfactory performance.

Yining Ma, Jingwen Li, Zhiguang Cao et al.

Recently, Transformer has become a prevailing deep architecture for solving vehicle routing problems (VRPs). However, it is less effective in learning improvement models for VRP because its positional encoding (PE) method is not suitable in representing VRP solutions. This paper presents a novel Dual-Aspect Collaborative Transformer (DACT) to learn embeddings for the node and positional features separately, instead of fusing them together as done in existing ones, so as to avoid potential noises and incompatible correlations. Moreover, the positional features are embedded through a novel cyclic positional encoding (CPE) method to allow Transformer to effectively capture the circularity and symmetry of VRP solutions (i.e., cyclic sequences). We train DACT using Proximal Policy Optimization and design a curriculum learning strategy for better sample efficiency. We apply DACT to solve the traveling salesman problem (TSP) and capacitated vehicle routing problem (CVRP). Results show that our DACT outperforms existing Transformer based improvement models, and exhibits much better generalization performance across different problem sizes on synthetic and benchmark instances, respectively.

Ru Zhang, Wenxin Sun, Jianyi Liu et al.

Advanced Persistent Threat (APT) attack, also known as directed threat attack, refers to the continuous and effective attack activities carried out by an organization on a specific object. They are covert, persistent and targeted, which are difficult to capture by traditional intrusion detection system(IDS). The traffic generated by the APT organization, which is the organization that launch the APT attack, has a high similarity, especially in the Command and Control(C2) stage. The addition of features for APT organizations can effectively improve the accuracy of traffic detection for APT attacks. This paper analyzes the DNS and TCP traffic of the APT attack, and constructs two new features, C2Load_fluct (response packet load fluctuation) and Bad_rate (bad packet rate). The analysis showed APT attacks have obvious statistical laws in these two features. This article combines two new features with common features to classify APT attack traffic. Aiming at the problem of data loss and boundary samples, we improve the Adaptive Synthetic(ADASYN) Sampling Approach and propose the PADASYN algorithm to achieve data balance. A traffic classification scheme is designed based on the AdaBoost algorithm. Experiments show that the classification accuracy of APT attack traffic is improved after adding new features to the two datasets so that 10 DNS features, 11 TCP and HTTP/HTTPS features are used to construct a Features set. On the two datasets, F1-score can reach above 0.98 and 0.94 respectively, which proves that the two new features in this paper are effective for APT traffic detection.