A complementary item is an item that pairs well with another item when consumed together. In the context of e-commerce, providing recommendations for complementary items is essential for both customers and stores. Current models for suggesting complementary items often rely heavily on user behavior data, such as co-purchase relationships. However, just because two items are frequently bought together does not necessarily mean they are truly complementary. Relying solely on co-purchase data may not align perfectly with the goal of making meaningful complementary recommendations. In this paper, we introduce the concept of "coherent complement recommendation", where "coherent" implies that recommended item pairs are compatible and relevant. Our approach builds upon complementary item pairs, with a focus on ensuring that recommended items are well used together and contextually relevant. To enhance the explainability and coherence of our complement recommendations, we fine-tune the Large Language Model (LLM) with coherent complement recommendation and explanation generation tasks since LLM has strong natural language explanation generation ability and multi-task fine-tuning enhances task understanding. Experimental results indicate that our model can provide more coherent complementary recommendations than existing state-of-the-art methods, and human evaluation validates that our approach achieves up to a 48% increase in the coherent rate of complement recommendations.

We present ACMI, an FM-based indoor localization system that does not require proactive site profiling. ACMI constructs the fingerprint database based on pure estimation of indoor received signal strength (RSS) distribution, where only the signals transmitted from commercial FM radio stations are used. Based on extensive field measurement study, we established our own signal propagation model that harnesses FM radio characteristics and open information of FM transmission towers in combination with the floor-plan of a building. Output of the model is an RSS fingerprint database. Using the fingerprint database as a knowledge base, ACMI refines a positioning result via the two-step process; parameter calibration and path matching, during its runtime. Without site profiling, our evaluation indicates that ACMI in seven campus locations and three downtown buildings using eight distinguished FM stations finds positions with only about 6 and 10 meters of errors on average, respectively.

Spectrum sensing, the task of discovering spectrum usage at a given location, is a fundamental problem in dynamic spectrum access networks. While sensing in narrow spectrum bands is well studied in previous work, wideband spectrum sensing is challenging since a wideband radio is generally too expensive and power consuming for mobile devices. Sequential scan, on the other hand, can be very slow if the wide spectrum band contains many narrow channels. In this paper, we propose an analog-filter based spectrum sensing technique, which is much faster than sequential scan and much cheaper than using a wideband radio. The key insight is that, if the sum of energy on a contiguous band is low, we can conclude that all channels in this band are clear with just one measurement. Based on this insight, we design an intelligent search algorithm to minimize the number of total measurements. We prove that the algorithm has the same asymptotic complexity as compressed sensing while our design is much simpler and easily implementable in the real hardware. We show the availability of our technique using hardware devices that include analog filters and analog energy detectors. Our extensive evaluation using real TV “white space” signals shows the effectiveness of our technique.

We present ACMI, an FM-based indoor localization that does not require proactive site profiling. ACMI constructs the fingerprint database based on the pure estimation of indoor RSS distribution, where the signals transmitted from commercial FM radio stations are used. For this, ACMI makes use of our signal model harnessing public transmission information of FM stations in a combination with a floorplan of a building. Using the fingerprint database as the knowledge base, ACMI actively performs multi-level online signal matching to infer the current location of a mobile user. ACMI achieves good indoor localization accuracy even without site profiling efforts. We evaluate ACMI with extensive indoor experiments in 7 different locations with over 1,100 indoor spots. The results show that ACMI achieves up to 89% room identification and accuracy of 6m localization error on average using 8 FM broadcast signals.

A practical protocol jointly considering PHY and MAC for MIMO based concurrent transmissions in wireless ad hoc networks, called Contrabass, is presented. Concurrent transmissions refer to simultaneous transmissions by multiple nodes over the same carrier frequency within the same interference range. Contrabass is the first-to-date open-loop based concurrent transmission protocol which implements simultaneous channel training for concurrently transmitting links without any control message exchange. Its MAC protocol is designed for each active transmitter to independently decide to transmit with near optimal transmission probability. Contrabass maximizes the number of successful concurrent transmissions, thus achieving very high aggregate throughput, low delays and scalability even under dynamic environments. The design choices of Contrabass are deliberately made to enable practical implementation which is demonstrated through GNURadio implementation and experimentation.

A PHY and MAC protocol for MIMO concurrent transmissions, called Contrabass, is presented. Concurrent transmissions, also referred to as multi-user MIMO, are simultaneous transmissions by multiple interfering nodes over the same carrier frequency. Concurrent transmissions technique has the potential of mitigating the overhead of MAC protocols by amortizing protocol overhead among multiple packets. However, existing proposals for concurrent transmissions could not achieve this as MIMO channel training and collision avoidance typically involve an expensive process of coordination and control message exchanges. This overhead has made MIMO concurrent transmission Impractical and thus unused in real applications. Contrabass implements simultaneous channel training and optimal transmission control without any coordination. As a result, Contrabass achieves very high aggregate throughput, low delays and scalability even under dynamic environments and outperforms the existing MIMO protocols. This is the first practical implementation of MIMO-based concurrent transmissions. We implemented Contrabass in GNU radios and also in NS-2.