Relation Extraction (RE) is a key task in table understanding, aiming to extract semantic relations between columns.However, complex tables with hierarchical headers are hard to obtain high-quality textual formats (e.g., Markdown) for input under practical scenarios like webpage screenshots and scanned documents, while table images are more accessible and intuitive.Besides, existing works overlook the need of mining relations among multiple columns rather than just the semantic relation between two specific columns in real-world practice.In this work, we explore utilizing Multimodal Large Language Models (MLLMs) to address RE in tables with complex structures.We creatively extend the concept of RE to include calculational relations, enabling multi-task learning of both semantic and calculational RE for mutual reinforcement.Specifically, we reconstruct table images into graph structure based on neighboring nodes to extract graph-level visual features.Such feature enhancement alleviates the insensitivity of MLLMs to the positional information within table images.We then propose a Chain-of-Thought distillation framework with self-correction mechanism to enhance MLLMs' reasoning capabilities without increasing parameter scale.Our method significantly outperforms most baselines on wide datasets.Additionally, we release a benchmark dataset for calculational RE in complex tables.

Weight-only quantization has been demonstrated to be an effective method to reduce memory and storage requirements for large language model (LLMs) deployed on resourceconstrained edge devices. However, it introduces frequent mixed precision operations between interger (INT) weights and floatingpoint (FP) activations. Conventional solutions typically quantize FP activations or dequantize INT weights during inference, which can degrade numerical accuracy and incur additional software overhead, ultimately slowing down inference. To eliminate this overhead, we propose VIFMM, an instruction set architecture (ISA) solution based on RISC-V vector (RVV) extension, supporting INT-FP mixed-precision computing, especially for operations between INT weights and FP activations, avoiding non-negligible data conversion in software. We implement VIFMM based on Ara, an open-source RVV core, along with a hardwarebased precision compensation mechanism to reuse the integer multiply accumulate calculation (MAC) units, to reduce hardware consumption and increase data precision. Experimental results show that VIFMM improves inference speed by 4%, with only 1.13% increase in hardware resource and 0.67% increase in power consumption, compared to conventional vector instructions. These findings illustrate that architecture-level support for INT-FP mixed-precision operations can significantly enhance LLM inference efficiency without sacrificing accuracy.

This paper presents a new vector quantization approach based on non-binary polar codes. Specifically, we use the non-binary polar code soft-input decoder as the encoder for vector quantization to realize source compression, while the non-binary polar encoder as the vector quantization decoder for source reconstruction. Further, we optimize the design of the reconstruction constellation set and the transition probability of the test channel. Our proposed scheme not only effectively mitigates the loss associated with the binary-to-multilevel mapping process, but is also adaptable to various sources with different distributions. Our scheme exhibits superior performance compared to the baseline binary polar coded method. Numerical evaluations under both continuous Gaussian and discrete multilevel source conditions demonstrate that the rate-distortion performance is close to the theoretical bound.

As a means of mass transportation, the safety and passengers’ comfort of maglev trains are paramount. Nevertheless, evaluating the two indices of maglev trains is challenging based on individual error measures. Of note, track irregularities on a maglev line may introduce unsafety concerns and discomfort for passengers. Finding a suitable control scheme that balances trade-offs is a complex issue. This paper presents a new model for chassis control to address this challenge, jointly considering secondary suspension dynamics (connecting chassis and carriage) and track irregularities. Then, we first develop an adaptive disturbance observer(ADO) and a novel safe backstepping control barrier function (CBF) to reject track irregularities and prevent physical contact (unsafety issue), and extend to address ride comfort. Moreover, the feasibility of the proposed method associated with the conflicts between safety and passenger comfort is analyzed. Finally, The trade-off between safety and ride comfort is mediated by a quadratic program (QP). A substantial amount of simulation results validate that the proposed control method addresses the ride comfort and safety of maglev trains well.

Matching tabular data to a knowledge graph (KG) is critical for understanding the semantic column types, column relationships, and entities of a table. Existing matching approaches rely heavily on core columns that represent primary subject entities on which other columns in the table depend. However, discovering these core columns before understanding the table’s semantics is challenging. Most prior works use heuristic rules, such as the leftmost column, to discover a single core column, while an insightful discovery of the core column set that accurately captures the dependencies between columns is often overlooked. To address these challenges, we introduce Dependency-aware Core Column Set Discovery ( DaCo ), an iterative method that uses a novel rough matching strategy to identify both inter-column dependencies and the core column set. Additionally, DaCo can be seamlessly integrated with pre-trained language models, as proposed in the optimization module. Unlike other methods, DaCo does not require labeled data or contextual information, making it suitable for real-world scenarios. In addition, it can identify multiple core columns within a table, which is common in real-world tables. We conduct experiments on six datasets, including five datasets with single core columns and one dataset with multiple core columns. Our experimental results show that DaCo outperforms existing core column set detection methods, further improving the effectiveness of table understanding tasks.

Unavoidable track irregularities continuously excite a traveling high-speed electromagnetic suspension (EMS) train. This issue deteriorates the levitation performances such as stability and comfortability. A robust levitation controller is critical for the EMS unit to suppress irregularities. In this work, a novel composite control scheme is proposed for high-speed maglev trains. It combines a fixed-time disturbance observer (FTDO) and a global finite-time controller (FTC). The FTDO is designed to estimate irregularities precisely within desired bandwidth, while the FTC is to address the fast dynamics of the levitation system of a high-speed maglev train. In addition, the command filter is integrated to increase the practicability of the composite control scheme. Theoretical analysis establishes the stability of the whole system. With ideal irregularity models and the field data measured from a commercial line, numerical simulations under various operation scenarios, including high speed, time-varying speed, gust and slope, verify that the proposed control scheme improves the air gap response while maintaining the stationarity of primary levitation and the ride comfortability for passengers. The simulations on a full bogie with load also demonstrate effectiveness of the proposed scheme.