NetSafe: Exploring the Topological Safety of Multi-agent Networks

Large language models (LLMs) have empowered nodes within multi-agent networks with intelligence, showing growing applications in both academia and industry. However, how to prevent these networks from generating malicious information remains unexplored with previous research on single LLM's safety be challenging to transfer. In this paper, we focus on the safety of multi-agent networks from a topological perspective, investigating which topological properties contribute to safer networks. To this end, we propose a general framework, NetSafe along with an iterative RelCom interaction to unify existing diverse LLM-based agent frameworks, laying the foundation for generalized topological safety research. We identify several critical phenomena when multi-agent networks are exposed to attacks involving misinformation, bias, and harmful information, termed as Agent Hallucination and Aggregation Safety. Furthermore, we find that highly connected networks are more susceptible to the spread of adversarial attacks, with task performance in a Star Graph Topology decreasing by 29.7%. Besides, our proposed static metrics aligned more closely with real-world dynamic evaluations than traditional graph-theoretic metrics, indicating that networks with greater average distances from attackers exhibit enhanced safety. In conclusion, our work introduces a new topological perspective on the safety of LLM-based multi-agent networks and discovers several unreported phenomena, paving the way for future research to explore the safety of such networks.

Task Consistent Prototype Learning for Incremental Few-shot Semantic Segmentation

Incremental Few-Shot Semantic Segmentation (iFSS) tackles a task that requires a model to continually expand its segmentation capability on novel classes using only a few annotated examples. Typical incremental approaches encounter a challenge that the objective of the base training phase (fitting base classes with sufficient instances) does not align with the incremental learning phase (rapidly adapting to new classes with less forgetting). This disconnect can result in suboptimal performance in the incremental setting. This study introduces a meta-learning-based prototype approach that encourages the model to learn how to adapt quickly while preserving previous knowledge. Concretely, we mimic the incremental evaluation protocol during the base training session by sampling a sequence of pseudo-incremental tasks. Each task in the simulated sequence is trained using a meta-objective to enable rapid adaptation without forgetting. To enhance discrimination among class prototypes, we introduce prototype space redistribution learning, which dynamically updates class prototypes to establish optimal inter-prototype boundaries within the prototype space. Extensive experiments on iFSS datasets built upon PASCAL and COCO benchmarks show the advanced performance of the proposed approach, offering valuable insights for addressing iFSS challenges.

Get Rid of Task Isolation: A Continuous Multi-task Spatio-Temporal Learning Framework

Spatiotemporal learning has become a pivotal technique to enable urban intelligence. Traditional spatiotemporal models mostly focus on a specific task by assuming a same distribution between training and testing sets. However, given that urban systems are usually dynamic, multi-sourced with imbalanced data distributions, current specific task-specific models fail to generalize to new urban conditions and adapt to new domains without explicitly modeling interdependencies across various dimensions and types of urban data. To this end, we argue that there is an essential to propose a Continuous Multi-task Spatio-Temporal learning framework (CMuST) to empower collective urban intelligence, which reforms the urban spatiotemporal learning from single-domain to cooperatively multi-dimensional and multi-task learning. Specifically, CMuST proposes a new multi-dimensional spatiotemporal interaction network (MSTI) to allow cross-interactions between context and main observations as well as self-interactions within spatial and temporal aspects to be exposed, which is also the core for capturing task-level commonality and personalization. To ensure continuous task learning, a novel Rolling Adaptation training scheme (RoAda) is devised, which not only preserves task uniqueness by constructing data summarization-driven task prompts, but also harnesses correlated patterns among tasks by iterative model behavior modeling. We further establish a benchmark of three cities for multi-task spatiotemporal learning, and empirically demonstrate the superiority of CMuST via extensive evaluations on these datasets. The impressive improvements on both few-shot streaming data and new domain tasks against existing SOAT methods are achieved. Code is available at https://github.com/DILab-USTCSZ/CMuST.

Caching Content Placement and Beamforming Co-design for IRS-Aided MIMO Systems with Imperfect CSI

When offloading links encounter deep fading and obstruction, edge caching cannot fully enhance wireless network performance and improve the QoS of edge nodes, as it fails to effectively reduce backhaul burden. The emerging technology of intelligent reflecting surfaces (IRS) compensates for this disadvantage by creating a smart and reconfigurable wireless environment. Subsequently, we jointly design content placement and active/passive beamforming to minimize network costs under imperfect channel state information (CSI) in the IRS-oriented edge caching system. This minimization problem is decomposed into two subproblems. The content placement subproblem is addressed by applying KKT optimality conditions. We then develop the alternating optimization method to resolve precoder and reflection beamforming. Specifically, we reduce transmission power by first fixing the phase shift, reducing the problem to a convex one relative to the precoder, which is solved through convex optimization. Next, we fix the precoder and resolve the resulting reflection beamforming problem using the penalty convex-concave procedure (CCP) method. Results demonstrate that our proposed method outperforms uniform caching and random phase approaches in reducing transmission power and saving network costs. Eventually, the proposed approach offers potential improvements in the caching optimization and transmission robustness of wireless communication with imperfect CSI.

ERCache: An Efficient and Reliable Caching Framework for Large-Scale User Representations in Meta's Ads System

The increasing complexity of deep learning models used for calculating user representations presents significant challenges, particularly with limited computational resources and strict service-level agreements (SLAs). Previous research efforts have focused on optimizing model inference but have overlooked a critical question: is it necessary to perform user model inference for every ad request in large-scale social networks? To address this question and these challenges, we first analyze user access patterns at Meta and find that most user model inferences occur within a short timeframe. T his observation reveals a triangular relationship among model complexity, embedding freshness, and service SLAs. Building on this insight, we designed, implemented, and evaluated ERCache, an efficient and robust caching framework for large-scale user representations in ads recommendation systems on social networks. ERCache categorizes cache into direct and failover types and applies customized settings and eviction policies for each model, effectively balancing model complexity, embedding freshness, and service SLAs, even considering the staleness introduced by caching. ERCache has been deployed at Meta for over six months, supporting more than 30 ranking models while efficiently conserving computational resources and complying with service SLA requirements.

ConServe: Harvesting GPUs for Low-Latency and High-Throughput Large Language Model Serving

Many applications are leveraging large language models (LLMs) for complex tasks, and they generally demand low inference latency and high serving throughput for interactive online jobs such as chatbots. However, the tight latency requirement and high load variance of applications pose challenges to serving systems in achieving high GPU utilization. Due to the high costs of scheduling and preemption, today's systems generally use separate clusters to serve online and offline inference tasks, and dedicate GPUs for online inferences to avoid interference. This approach leads to underutilized GPUs because one must reserve enough GPU resources for the peak expected load, even if the average load is low. This paper proposes to harvest stranded GPU resources for offline LLM inference tasks such as document summarization and LLM benchmarking. Unlike online inferences, these tasks usually run in a batch-processing manner with loose latency requirements, making them a good fit for stranded resources that are only available shortly. To enable safe and efficient GPU harvesting without interfering with online tasks, we built ConServe, an LLM serving system that contains (1) an execution engine that preempts running offline tasks upon the arrival of online tasks, (2) an incremental checkpointing mechanism that minimizes the amount of recomputation required by preemptions, and (3) a scheduler that adaptively batches offline tasks for higher GPU utilization. Our evaluation demonstrates that ConServe achieves strong performance isolation when co-serving online and offline tasks but at a much higher GPU utilization. When colocating practical online and offline workloads on popular models such as Llama-2-7B, ConServe achieves 2.35$\times$ higher throughput than state-of-the-art online serving systems and reduces serving latency by 84$\times$ compared to existing co-serving systems.

From Facts to Insights: A Study on the Generation and Evaluation of Analytical Reports for Deciphering Earnings Calls

This paper explores the use of Large Language Models (LLMs) in the generation and evaluation of analytical reports derived from Earnings Calls (ECs). Addressing a current gap in research, we explore the generation of analytical reports with LLMs in a multi-agent framework, designing specialized agents that introduce diverse viewpoints and desirable topics of analysis into the report generation process. Through multiple analyses, we examine the alignment between generated and human-written reports and the impact of both individual and collective agents. Our findings suggest that the introduction of additional agents results in more insightful reports, although reports generated by human experts remain preferred in the majority of cases. Finally, we address the challenging issue of report evaluation, we examine the limitations and strengths of LLMs in assessing the quality of generated reports in different settings, revealing a significant correlation with human experts across multiple dimensions.

Learning Frame-Wise Emotion Intensity for Audio-Driven Talking-Head Generation

Human emotional expression is inherently dynamic, complex, and fluid, characterized by smooth transitions in intensity throughout verbal communication. However, the modeling of such intensity fluctuations has been largely overlooked by previous audio-driven talking-head generation methods, which often results in static emotional outputs. In this paper, we explore how emotion intensity fluctuates during speech, proposing a method for capturing and generating these subtle shifts for talking-head generation. Specifically, we develop a talking-head framework that is capable of generating a variety of emotions with precise control over intensity levels. This is achieved by learning a continuous emotion latent space, where emotion types are encoded within latent orientations and emotion intensity is reflected in latent norms. In addition, to capture the dynamic intensity fluctuations, we adopt an audio-to-intensity predictor by considering the speaking tone that reflects the intensity. The training signals for this predictor are obtained through our emotion-agnostic intensity pseudo-labeling method without the need of frame-wise intensity labeling. Extensive experiments and analyses validate the effectiveness of our proposed method in accurately capturing and reproducing emotion intensity fluctuations in talking-head generation, thereby significantly enhancing the expressiveness and realism of the generated outputs.

VPTQ: Extreme Low-bit Vector Post-Training Quantization for Large Language Models

Scaling model size significantly challenges the deployment and inference of Large Language Models (LLMs). Due to the redundancy in LLM weights, recent research has focused on pushing weight-only quantization to extremely low-bit (even down to 2 bits). It reduces memory requirements, optimizes storage costs, and decreases memory bandwidth needs during inference. However, due to numerical representation limitations, traditional scalar-based weight quantization struggles to achieve such extreme low-bit. Recent research on Vector Quantization (VQ) for LLMs has demonstrated the potential for extremely low-bit model quantization by compressing vectors into indices using lookup tables. In this paper, we introduce Vector Post-Training Quantization (VPTQ) for extremely low-bit quantization of LLMs. We use Second-Order Optimization to formulate the LLM VQ problem and guide our quantization algorithm design by solving the optimization. We further refine the weights using Channel-Independent Second-Order Optimization for a granular VQ. In addition, by decomposing the optimization problem, we propose a brief and effective codebook initialization algorithm. We also extend VPTQ to support residual and outlier quantization, which enhances model accuracy and further compresses the model. Our experimental results show that VPTQ reduces model quantization perplexity by $0.01$-$0.34$ on LLaMA-2, $0.38$-$0.68$ on Mistral-7B, $4.41$-$7.34$ on LLaMA-3 over SOTA at 2-bit, with an average accuracy improvement of $0.79$-$1.5\%$ on LLaMA-2, $1\%$ on Mistral-7B, $11$-$22\%$ on LLaMA-3 on QA tasks on average. We only utilize $10.4$-$18.6\%$ of the quantization algorithm execution time, resulting in a $1.6$-$1.8\times$ increase in inference throughput compared to SOTA.

One Model for Two Tasks: Cooperatively Recognizing and Recovering Low-Resolution Scene Text Images by Iterative Mutual Guidance

Scene text recognition (STR) from high-resolution (HR) images has been significantly successful, however text reading on low-resolution (LR) images is still challenging due to insufficient visual information. Therefore, recently many scene text image super-resolution (STISR) models have been proposed to generate super-resolution (SR) images for the LR ones, then STR is done on the SR images, which thus boosts recognition performance. Nevertheless, these methods have two major weaknesses. On the one hand, STISR approaches may generate imperfect or even erroneous SR images, which mislead the subsequent recognition of STR models. On the other hand, as the STISR and STR models are jointly optimized, to pursue high recognition accuracy, the fidelity of SR images may be spoiled. As a result, neither the recognition performance nor the fidelity of STISR models are desirable. Then, can we achieve both high recognition performance and good fidelity? To this end, in this paper we propose a novel method called IMAGE (the abbreviation of Iterative MutuAl GuidancE) to effectively recognize and recover LR scene text images simultaneously. Concretely, IMAGE consists of a specialized STR model for recognition and a tailored STISR model to recover LR images, which are optimized separately. And we develop an iterative mutual guidance mechanism, with which the STR model provides high-level semantic information as clue to the STISR model for better super-resolution, meanwhile the STISR model offers essential low-level pixel clue to the STR model for more accurate recognition. Extensive experiments on two LR datasets demonstrate the superiority of our method over the existing works on both recognition performance and super-resolution fidelity.