Large language models (LLMs) have exhibited remarkable few-shot learning capabilities and unified the paradigm of NLP tasks through the in-context learning (ICL) technique. Despite the success of ICL, the quality of the exemplar demonstrations can significantly influence the LLM’s performance. Existing exemplar selection methods mainly focus on the semantic similarity between queries and candidate exemplars. On the other hand, the logical connections between reasoning steps can also be beneficial to depict the problem-solving process. This paper proposes a novel method named Reasoning Graph-enhanced Exemplar Retrieval (RGER). RGER first queries LLM to generate an initial response and then expresses intermediate problem-solving steps to a graph structure. After that, it employs a graph kernel to select exemplars with semantic and structural similarity. Extensive experiments demonstrate the structural relationship is helpful to the alignment of queries and candidate exemplars. The efficacy of RGER on mathematics and logical reasoning tasks showcases its superiority over state-of-the-art retrieval-based approaches.

Cross-modal hashing (CMH) has been widely used in multimedia retrieval applications for its low storage cost and fast indexing speed. Thanks to the success of deep learning, cross-modal hashing has made significant progress with high-quality deep features. However, the modal gap is still a crucial bottleneck for existing cross-modal hashing methods: the commonly used convolutional neural network and bag-of-words encoders are customized for single modal prior, limiting the models to learn semantics representation in a cross-modal space. To overcome modality heterogeneity, we propose a shared transformer encoder (UniHash) to unify the cross-modal hashing into the same semantic space. A contrastive label correlation learning (CLC) loss using the category labels as modality bridge is designed together to improve the representation quality. Moreover, we take advantage of the multi-hot label space and propose a negative label generation (NegLG) strategy to get richer and uniformly distributed negative labels for contrast. Extensive experiments on three benchmarks verify the advantage of our proposed method. Besides, the proposed UniHash outperforms state-of-the-art cross-modal hashing methods significantly, establishing a new important baseline for the cross-modal hashing research. Codes are released github.com/idealwhite/Unihash.

This paper presents an approach to Alzheimer’s disease (AD) diagnosis from spontaneous speech using an end-to-end ASR-enhanced neural network. Under the condition that only audio data are provided and accurate transcripts are unavailable, this paper proposes a system that can analyze utterances to differentiate between AD patients, healthy controls, and individuals with mild cognitive impairment. The ASR-enhanced model comprises automatic speech recognition (ASR) with an encoder-decoder structure and the encoder followed by an AD classification network. The encoder takes a Mel spectrogram as input and transforms it into high-level acoustic features that correlate with AD. The classification network then maps intermediate acoustic features to three categories. In the training phase, the AD classification and speech recognition tasks are optimized simultaneously. Experimental results obtained from an AD recognition dataset of Chinese spontaneous speech1illustrate the effectiveness of integrating ASR into AD diagnosis in an end-to-end manner. Further, our model has low dependency on accurate ASR transcripts. This work achieved accuracy scores of 89.1% and 82.6% for long and short utterance tracks, respectively.

Integrating visual features has been proved useful for natural language understanding tasks. Nevertheless, most existing multimodal language models highly rely on training on aligned image and text data. In this paper, we propose a novel semi-supervised visual integration framework for pre-trained language models. In the framework, the visual features are obtained through a sentence visualization and vision-language fusion mechanism. The uniqueness includes: 1) the integration is conducted via a semi-supervised framework and does not require aligned images for the processed sentences. 2) the framework works as an auxiliary component, and will not affect the language processing ability of the integrated language model. Experimental results on both natural language inference and reading comprehension tasks demonstrate that our framework improves the strong baseline language models. Considering that our framework only requires an image database, and does not require aligned images for the processed texts, it provides a feasible way for multimodal language learning.

Recently with the sharing economy of WiFi, IEEE 802.11 Wireless Mesh Networks (WMNs) have been deployed exponentially in more places. With the booming increase of density of APs as well as mobile clients (MCs), the handoff between MCs and APs becomes more frequent than ever before. One problem of most existing network-assisted fast handoff mechanisms is that an AP needs to switch to different channel to probe MCs, thus disrupting its current communications. In this paper, we propose a fast handoff mechanism called OppoScan. OppoScan opportunistically leverages nearby MCs and APs to produce the required information of neighboring AP for handoff, thus significantly decrease the number of switching channel of APs. We have implemented the OppoScan in about 200 APs in two shopping malls and the results demonstrated the performance of the proposed scheme.

Pervasive computing environments (PvCE) are embedded with interconnected smart devices which provide users with services desired. To meet requirements of users, smart devices with different kinds of functions may need to be associated together to provide the service described in the user requirement, which is called service composition. As the service composition environment may be dynamic and large scale, centralized service composition algorithm is usually inefficient due to message cost. On the other hand, a decentralized approach, which employs pre-determined coordinators to search and compose service, may have high cost as well. In this paper, we discuss a localized approach for service composition based on the Ubiquitous Interacting Object (UIO) model we have proposed earlier. UIO is an abstraction of physical devices in PvCE with ability to find and collaborate with other devices through exposing their capabilities as services. In our localized service composition algorithm (LASEC), UIOs collaborate with each other in a bottom-up, localized manner to compose required service without requiring global knowledge. To solve the problem of blind compositions in LASEC, we propose a novel mechanism called Alien-information-based Acknowledging (A-Ack), in which a UIO decides on collaborating with another UIO only after obtaining some additional information from the collaboration candidate. Specifically, this information refers to ability of a given UIO to compose another part of the service. Proposed LASEC is message-efficient and quality-guaranteed. Extensive simulations of LASEC as well as existing decentralized and pull-based centralized algorithms have been conducted. The results show the relatively low communication cost and composition time of LASEC. Moreover, we demonstrate feasibility of our approach with a prototype implementation.

Cooperative forwarding has been considered as an effective strategy for improving the geographic routing performance in wireless sensor networks (WSNs). However, we observe that existing works do not fully utilize the forwarding opportunities provided by available neighboring nodes. In this paper, we redesign the cooperative forwarding process and present a novel cooperative contention-based forwarding (CCBF) protocol for WSNs. CCBF extends the scope of cooperation and attains the full potential of cooperative forwarding at the expense of sending one additional control message on demand. We conduct extensive simulations to study the performance of the proposed CCBF compared with existing protocols. Simulation results demonstrate that CCBF decreases the end-to-end hop counts, hence, further improves the end-to-end energy efficiency and latency. Remarkably, it provides up to 50% improvement for the packet loss ratio when the retransmission mechanism is not adopted.

In a pervasive computing environment, the devices are embedded in the physical world, providing services and interconnected by a communication network. Composition of these services is important issue of pervasive applications which integrate the physical and cyber worlds. Most existing research on service composition in pervasive computing relies on the existence of one or more entities that maintain the global service information. However, such an approach is not always practical due to dynamicity of the environment. In this paper, we propose a fully decentralized approach to service composition. We first model the service composition problem as finding an overlay of the communication network that matches the composition graph. The problem is proved to be NP-complete. We propose an algorithm for the devices to cooperatively construct the requested services through localized interactions. For the purpose of reducing redundant broadcast we propose the service composition backbone built in a fully localized way. We have carried out extensive simulations. to evaluate the performance of our algorithm. Compared with existing pull-based centralized techniques our decentralized service composition algorithm on the service composition backbone is more efficient in terms of response delay and message overhead, while achieving similar quality of composed service.