Structured pruning offers a hardware-friendly approach for efficient LLM inference.Early static methods determine fixed subnetworks through offline calibration, suffering from performance degradation and calibration sensitivity.Recent methods explore input-adaptive pruning by selecting a subset of tokens as probes to estimate hidden activations for online pruning decisions.However, existing probe selection strategies fail to identify outlier-triggering tokens, and uniform layerwise sparsity misaligns with heterogeneous outlier distributions, leading to critical channels being incorrectly pruned.Therefore, we propose OCP (Outlier-Centric Probing for structured pruning), a principled framework that prioritizes capturing outlier-triggering tokens rather than reconstructing full hidden distributions.Specifically, OCP includes three key components: (1) sensitivity-weighted probing for FFN layers that identifies outlier patterns via precomputed weight aggregations, (2) attention-accumulated probing that leverages preceding attention matrices to identify salient tokens, and (3) online adaptive sparsity allocation that dynamically adjusts layerwise pruning based on history-guided outlier statistics.Extensive experiments on LLaMA2, LLaMA3, and OPT demonstrate that OCP consistently outperforms state-of-the-art methods across benchmarks, achieving up to 25% perplexity reduction at 1.6× speedup.

Neural networks have achieved remarkable success across various fields. However, the lack of interpretability limits their practical use, particularly in critical decision-making scenarios. Posthoc interpretability, which provides explanations for pretrained models, is often at risk of fidelity and robustness. This has inspired a rising interest in self-interpretable neural networks (SINNs), which inherently reveal the prediction rationale through model structures. Despite this progress, existing research remains fragmented, relying on intuitive designs tailored to specific tasks. To bridge these efforts and foster a unified framework, we first collect and review existing works on SINNs and provide a structured summary of their methodologies from five key perspectives: attribution-based, function-based, concept-based, prototype-based, and rule-based self-interpretation. We also present concrete, visualized examples of model explanations and discuss their applicability across diverse scenarios, including image, text, graph data, and deep reinforcement learning (DRL). Additionally, we summarize existing evaluation metrics for self-interpretation and identify open challenges in this field, offering insights for future research. To support ongoing developments, we present a publicly accessible resource to track advancements in this domain: https://github.com/yangji721/Awesome-Self-Interpretable-Neural-Network

Continuously learning new skills is essential for talents to gain a competitive advantage in the labor market. Despite extensive efforts on relevance- or preference-based skill recommendations, little attention has been given to the practical effects of job skills in the market. To bridge this gap, we propose an explainable personalized skill learning recommendation system that considers the long-term learning benefits and costs. Specifically, we model skill learning utilities based on salary and learning cost associated with job positions and propose a multi-objective deep reinforcement learning framework to model and maximize long-term utilities. Furthermore, we propose a Self-explaining Skill Recommendation Deep Q-network (SeSRDQN) that captures and prototypes prevalent skill sets in the market into representative exemplars for decision-making. SeSRDQN quantitatively decomposes the talent’s long-term learning utility into contributions from each exemplar, offering a comprehensive and multi-factorial explanation across various skill learning options. To tackle the combinatorial complexity of the skill space, we develop an MCTS-based optimization-decoding iterative training procedure for explanation fidelity and human understandability. In this way, talents will receive a tailored roadmap of essential skills, complemented by exemplar-based explanations, to effectively plan their careers. Extensive experiments on a real-world dataset validate the effectiveness and explainability of our approach.

Light clients have been widely used in blockchain systems to support lightweight nodes by synchronizing and verifying block headers only. However, there are two major limitations with the current light client design. First, with the ever increasing blockchain size, the cost for light clients to process and store all the block headers would soon become prohibitively high. Second, only simple queries can be supported by light clients due to the limited functionality of block headers. To address these issues, in this paper, we propose DCert, a novel decentralized certification framework, to enable superlight clients with constant storage and state validation costs. The main idea is to leverage a trusted enclave (e.g., Intel SGX) to recursively certify the entire history of the blockchain. With DCert, the blockchain integrity can be easily validated by superlight clients with a secure certificate. Furthermore, to support rich verifiable queries on light clients, DCert can be extended to certify authenticated indexes for different types of queries on an as-needed basis. While DCert is compatible with existing blockchain systems, its security is guaranteed by the trusted enclave. Our benchmark-based empirical study shows that DCert incurs a small certification overhead, yet it is capable of supporting efficient verifiable queries with a constant storage size of 2.97 KB and a constant bootstrapping time of 0.14 ms.

We demonstrate vCBIR, a verifiable search engine for Content-Based Image Retrieval. vCBIR allows a small or medium-sized enterprise to outsource its image database to a cloud-based service provider and ensures the integrity of query processing. Like other common data-as-a-service (DaaS) systems, vCBIR consists of three parties: (i) the image owner who outsources its database, (ii) the service provider who executes the authenticated query processing, and (iii) the client who issues search queries. By employing a novel query authentication scheme proposed in our prior work [4], the system not only supports cloud-based image retrieval, but also generates a cryptographic proof for each query, by which the client could verify the integrity of query results. During the demonstration, we will showcase the usage of vCBIR and also provide attendees interactive experience of verifying query results against an untrustworthy service provider through graphical user interface (GUI).

With the proliferation of cloud computing and data-as-a-service (DaaS), more and more organizations and individuals outsource their data to a third-party service provider. While enjoying the benefits of cloud-based data outsourcing, the data owners are at the risk of losing control of data integrity and access management. In this demonstration, we present a system called vABS, which enables verifiable Attribute-Based Search over shared cloud data. The vABS system adopts the common DaaS architecture, in which the server provides search services to users on behalf of data owners. By employing a novel zero-knowledge approach proposed in our prior work [1], vABS not only provides users with good search experiences, but also supports authenticated query processing with fine-grained access control, which is crucial to many high-security applications.