arXiv subjects
Publications and source records attributed to Hao Guo.
We establish a relationship between a certain notion of covering complexity of a Riemannian spin manifold and positive lower bounds on its scalar curvature. This makes use of a pairing between quantitative operator $K$-theory and Lipschitz topological $K$-theory, combined with an earlier vanishing theorem for the quantitative higher index.
We first compare the geometric frameworks behind the Uhlmann and Berry phases in a fiber-bundle language and then evaluate the Uhlmann phases of bosonic and fermionic coherent states. The Uhlmann phases of both coherent states are shown to carry geometric information and decrease smoothly with temperature. Importantly, the Uhlmann phases approach the corresponding Berry phases as temperature decreases. Together with previous examples in the literature, we propose a correspondence between the Uhlmann and Berry phases in the zero-temperature limit as a general property except some special cases and present a conditional proof of the correspondence.
Network-controlled repeater (NCR) has been recently considered as a study-item in 3GPP Release 18, and the discussions are continuing in a work-item. In this paper, we introduce the concept of NCRs, as a possible low-complexity device to support for network densification and compare the performance of the NCRs with those achieved by reconfigurable intelligent surfaces (RISs). The results are presented for the cases with different beamforming methods and hardware impairment models of the RIS. Moreover, we introduce the objectives of the 3GPP Release 18 NCR work-item and study the effect of different parameters on the performance of NCR-assisted networks. As we show, with a proper deployment, the presence of NCRs/RISs can improve the network performance considerably.
Integrated access and backhaul (IAB) is one of the promising techniques for 5G networks and beyond (6G), in which the same node/hardware is used to provide both backhaul and cellular services in a multi-hop fashion. Due to the sensitivity of the backhaul links with high rate/reliability demands, proper network planning is needed to make the IAB network performing appropriately and as good as possible. In this paper, we study the effect of deployment optimization on the coverage of IAB networks. We concentrate on the cases where, due to either geographical or interference management limitations, unconstrained IAB node placement is not feasible in some areas. To that end, we propose various millimeter wave (mmWave) blocking-aware constrained deployment optimization approaches. Our results indicate that, even with limitations on deployment optimization, network planning boosts the coverage of IAB networks considerably.
On the one hand, there is a growing demand for high throughput which can be satisfied thanks to the deployment of new networks using massive multiple-input multiple-output (MIMO) and beamforming. On the other hand, in some countries or cities, there is a demand for arbitrarily low electromagnetic field exposure (EMFE) of people not concerned by the ongoing communication, which slows down the deployment of new networks. Recently, it has been proposed to take the opportunity, when designing the future 6th generation (6G), to offer, in addition to higher throughput, a new type of service: arbitrarily low EMFE. Recent works have shown that a reconfigurable intelligent surface (RIS), jointly optimized with the base station (BS) beamforming can improve the received throughput at the desired location whilst reducing EMFE everywhere. In this paper, we introduce a new concept of a non-intended user (NIU). An NIU is a user of the network who requests low EMFE when he/she is not downloading/uploading data. An NIU lets his/her device, called NIU equipment (NIUE), exchange some control signaling and pilots with the network, to help the network avoid exposing NIU to waves that are transporting data for another user of the network: the intended user (IU), whose device is called IU equipment (IUE). Specifically, we propose several new schemes to maximize the IU throughput under an EMFE constraint at the NIU (in practice, an interference constraint at the NIUE). Several propagation scenarios are investigated. Analytical and numerical results show that proper power allocation and beam optimization can remarkably boost the EMFE-constrained system's performance with limited complexity and channel information.
We develop two connections between the quantitative framework of operator $K$-theory for geometric $C^*$-algebras and the problem of positive scalar curvature. First, we introduce a quantitative notion of higher index and use it to give a refinement of the well-known obstruction of Rosenberg to positive scalar curvature on closed spin manifolds coming from the higher index of the Dirac operator. We show that on a manifold with uniformly positive scalar curvature, the propagation at which the index of the Dirac operator vanishes is related inversely to the curvature lower bound. Second, we give an approach, using related techniques, to Gromov's band width conjecture, which has been the subject of recent work by Zeidler and Cecchini from a different point of view.
Averaging neural network weights sampled by a backbone stochastic gradient descent (SGD) is a simple yet effective approach to assist the backbone SGD in finding better optima, in terms of generalization. From a statistical perspective, weight averaging (WA) contributes to variance reduction. Recently, a well-established stochastic weight averaging (SWA) method is proposed, which is featured by the application of a cyclical or high constant (CHC) learning rate schedule (LRS) in generating weight samples for WA. Then a new insight on WA appears, which states that WA helps to discover wider optima and then leads to better generalization. We conduct extensive experimental studies for SWA, involving a dozen modern DNN model structures and a dozen benchmark open-source image, graph, and text datasets. We disentangle contributions of the WA operation and the CHC LRS for SWA, showing that the WA operation in SWA still contributes to variance reduction but does not always lead to wide optima. The experimental results indicate that there are global scale geometric structures in the DNN loss landscape. We then present an algorithm termed periodic SWA (PSWA) which makes use of a series of WA operations to discover the global geometric structures. PSWA outperforms its backbone SGD remarkably, providing experimental evidences for the existence of global geometric structures. Codes for reproducing the experimental results are available at https://github.com/ZJLAB-AMMI/PSWA.
Supply Chain Platforms (SCPs) provide downstream industries with numerous raw materials. Compared with traditional e-commerce platforms, data in SCPs is more sparse due to limited user interests. To tackle the data sparsity problem, one can apply Cross-Domain Recommendation (CDR) which improves the recommendation performance of the target domain with the source domain information. However, applying CDR to SCPs directly ignores the hierarchical structure of commodities in SCPs, which reduce the recommendation performance. To leverage this feature, in this paper, we take the catering platform as an example and propose GReS, a graphical cross-domain recommendation model. The model first constructs a tree-shaped graph to represent the hierarchy of different nodes of dishes and ingredients, and then applies our proposed Tree2vec method combining GCN and BERT models to embed the graph for recommendations. Experimental results on a commercial dataset show that GReS significantly outperforms state-of-the-art methods in Cross-Domain Recommendation for Supply Chain Platforms.
For a finitely generated discrete group $Γ$ acting properly on a spin manifold $M$, we formulate new topological obstructions to $Γ$-invariant metrics of positive scalar curvature on $M$ that take into account the cohomology of the classifying space $\underline{B}Γ$ for proper actions. In the cocompact case, this leads to a natural generalisation of Gromov-Lawson's notion of higher $\hat{A}$-genera to the setting of proper actions by groups with torsion. It is conjectured that these invariants obstruct the existence of $Γ$-invariant positive scalar curvature on $M$. For classes arising from the subring of $H^*(\underline{B}Γ,\mathbb{R})$ generated by elements of degree at most $2$, we are able to prove this, under suitable assumptions, using index-theoretic methods for projectively invariant Dirac operators and a twisted $L^2$-Lefschetz fixed-point theorem involving a weighted trace on conjugacy classes. The latter generalises a result of Wang-Wang to the projective setting. In the special case of free actions and the trivial conjugacy class, this reduces to a theorem of Mathai, which provided a partial answer to a conjecture of Gromov-Lawson on higher $\hat{A}$-genera. If $M$ is non-cocompact, we obtain obstructions to $M$ being a partitioning hypersurface inside a non-cocompact $Γ$-manifold with non-negative scalar curvature that is positive in a neighbourhood of the hypersurface. Finally, we define a quantitative version of the twisted higher index and use it to prove a parameterised vanishing theorem in terms of the lower bound of the total curvature term in the square of the twisted Dirac operator.
Reinforcement learning (RL) is a machine learning paradigm where an autonomous agent learns to make an optimal sequence of decisions by interacting with the underlying environment. The promise demonstrated by RL-guided workflows in unraveling electronic design automation problems has encouraged hardware security researchers to utilize autonomous RL agents in solving domain-specific problems. From the perspective of hardware security, such autonomous agents are appealing as they can generate optimal actions in an unknown adversarial environment. On the other hand, the continued globalization of the integrated circuit supply chain has forced chip fabrication to off-shore, untrustworthy entities, leading to increased concerns about the security of the hardware. Furthermore, the unknown adversarial environment and increasing design complexity make it challenging for defenders to detect subtle modifications made by attackers (a.k.a. hardware Trojans). In this brief, we outline the development of RL agents in detecting hardware Trojans, one of the most challenging hardware security problems. Additionally, we outline potential opportunities and enlist the challenges of applying RL to solve hardware security problems.
Stealthy hardware Trojans (HTs) inserted during the fabrication of integrated circuits can bypass the security of critical infrastructures. Although researchers have proposed many techniques to detect HTs, several limitations exist, including: (i) a low success rate, (ii) high algorithmic complexity, and (iii) a large number of test patterns. Furthermore, the most pertinent drawback of prior detection techniques stems from an incorrect evaluation methodology, i.e., they assume that an adversary inserts HTs randomly. Such inappropriate adversarial assumptions enable detection techniques to claim high HT detection accuracy, leading to a "false sense of security." Unfortunately, to the best of our knowledge, despite more than a decade of research on detecting HTs inserted during fabrication, there have been no concerted efforts to perform a systematic evaluation of HT detection techniques. In this paper, we play the role of a realistic adversary and question the efficacy of HT detection techniques by developing an automated, scalable, and practical attack framework, ATTRITION, using reinforcement learning (RL). ATTRITION evades eight detection techniques across two HT detection categories, showcasing its agnostic behavior. ATTRITION achieves average attack success rates of $47\times$ and $211\times$ compared to randomly inserted HTs against state-of-the-art HT detection techniques. We demonstrate ATTRITION's ability to evade detection techniques by evaluating designs ranging from the widely-used academic suites to larger designs such as the open-source MIPS and mor1kx processors to AES and a GPS module. Additionally, we showcase the impact of ATTRITION-generated HTs through two case studies (privilege escalation and kill switch) on the mor1kx processor. We envision that our work, along with our released HT benchmarks and models, fosters the development of better HT detection techniques.
Insertion of hardware Trojans (HTs) in integrated circuits is a pernicious threat. Since HTs are activated under rare trigger conditions, detecting them using random logic simulations is infeasible. In this work, we design a reinforcement learning (RL) agent that circumvents the exponential search space and returns a minimal set of patterns that is most likely to detect HTs. Experimental results on a variety of benchmarks demonstrate the efficacy and scalability of our RL agent, which obtains a significant reduction ($169\times$) in the number of test patterns required while maintaining or improving coverage ($95.75\%$) compared to the state-of-the-art techniques.
Two representations of mixed states by state-vectors, known as purified state and thermal vacuum, have been realized on quantum computers. While the two representations look similar, they differ by a partial transposition in the ancilla space. While ordinary observables cannot discern the two representations, we generalize the Berry phase of pure quantum states to mixed states and construct two geometric phases that can reflect the partial transposition. By generalizing the adiabatic condition, we construct the thermal Berry phase, whose values from the two representations can be different, However, the thermal Berry phase may contain non-geometrical contributions. Alternatively, we generalize the parallel-transport condition to include the system and ancilla and show the dynamical phase is excluded under parallel transport. The geometrical phase accumulated in parallel transport is the generalized Berry phase, which may or may not differentiate a purified state from a thermal vacuum depending on the protocol. The generalizations of the Berry phase to mixed states may be realized and measured on quantum computers via the two representations to reveal the rich physics of finite-temperature quantum systems.
Recent non-local self-attention methods have proven to be effective in capturing long-range dependencies for semantic segmentation. These methods usually form a similarity map of RC*C (by compressing spatial dimensions) or RHW*HW (by compressing channels) to describe the feature relations along either channel or spatial dimensions, where C is the number of channels, H and W are the spatial dimensions of the input feature map. However, such practices tend to condense feature dependencies along the other dimensions,hence causing attention missing, which might lead to inferior results for small/thin categories or inconsistent segmentation inside large objects. To address this problem, we propose anew approach, namely Fully Attentional Network (FLANet),to encode both spatial and channel attentions in a single similarity map while maintaining high computational efficiency. Specifically, for each channel map, our FLANet can harvest feature responses from all other channel maps, and the associated spatial positions as well, through a novel fully attentional module. Our new method has achieved state-of-the-art performance on three challenging semantic segmentation datasets,i.e., 83.6%, 46.99%, and 88.5% on the Cityscapes test set,the ADE20K validation set, and the PASCAL VOC test set,respectively.
By deriving a general framework and analyzing concrete examples, we demonstrate a class of dynamical quantum phase transitions (DQPTs) in one-dimensional two-band systems going through double-quench processes. When this type of DQPT occurs, the Loschmidt amplitude vanishes and the rate function remains singular after the second quench, meaning the final state continually has no overlap with the initial state. This type of DQPT is named metamorphic DQPT to differentiate it from ordinary DQPTs that only exhibit zero Loschmidt amplitude and singular rate function at discrete time points. The metamorphic DQPTs occur at zero as well as finite temperatures. Our examples of the Su-Schrieffer-Heeger (SSH) model and Kitaev chain illustrate the conditions and behavior of the metamorphic DQPT. Since ordinary DQPTs have been experimentally realized in many systems, similar setups with double quenches will demonstrate the metamorphic DQPT. Our findings thus provide additional controls of dynamical evolution of quantum systems.
In future wireless networks, one of the use-cases of interest is Internet-of-vehicles (IoV). Here, IoV refers to two different functionalities, namely, serving the in-vehicle users and supporting the connected-vehicle functionalities, where both can be well provided by the transceivers installed on top of vehicles. Such dual functionality of on-vehicle transceivers implies strict rate and reliability requirements, for which one may need to communicate at millimeter wave (mmW) frequencies. However, IoV communication at mmW requires up-to-date channel state information (CSI) and blockage avoidance. In this article, we incorporate the recently proposed concept of predictor antennas (PAs) into a large-scale cooperative PA (LSCPA) setup where both temporal blockages and CSI out-dating are avoided via base stations (BSs)/vehicles cooperation. Summarizing the ongoing standardization progress enabling IoV communications, we present the potentials and challenges of the LSCPA setup, and compare the effect of cooperative and non-cooperative schemes on the performance of IoV links. As we show, BSs cooperation and blockage/CSI prediction can boost the performance of IoV links remarkably.
We present a theory of a two-component atomic Fermi gas with tunable attractive contact interactions on a spherical shell going through the Bardeen-Cooper-Schrieffer (BCS) - Bose Einstein condensation (BEC) crossover, inspired by the realizations of spherical bubble traps for ultracold atoms in microgravity. The derivation follows the BCS-Leggett theory to obtain the gap and number equations. The BCS-BEC crossover can be induced by tuning the interaction, and the properly normalized gap and chemical potential exhibit universal behavior regardless of the planar or spherical geometry. Nevertheless, the spherical-shell geometry introduces another way of inducing the crossover by the curvature. The curvature-induced BCS-BEC crossover is made possible by fixing the particle number and interaction strength while shrinking the sphere, causing a reduction to the ratio of the pairing and kinetic energies and pushing the system towards the BCS limit. The saturation of the superfluid density further confirms the ground state is a Fermi superfluid.
The non-local network has become a widely used technique for semantic segmentation, which computes an attention map to measure the relationships of each pixel pair. However, most of the current popular non-local models tend to ignore the phenomenon that the calculated attention map appears to be very noisy, containing inter-class and intra-class inconsistencies, which lowers the accuracy and reliability of the non-local methods. In this paper, we figuratively denote these inconsistencies as attention noises and explore the solutions to denoise them. Specifically, we inventively propose a Denoised Non-Local Network (Denoised NL), which consists of two primary modules, i.e., the Global Rectifying (GR) block and the Local Retention (LR) block, to eliminate the inter-class and intra-class noises respectively. First, GR adopts the class-level predictions to capture a binary map to distinguish whether the selected two pixels belong to the same category. Second, LR captures the ignored local dependencies and further uses them to rectify the unwanted hollows in the attention map. The experimental results on two challenging semantic segmentation datasets demonstrate the superior performance of our model. Without any external training data, our proposed Denoised NL can achieve the state-of-the-art performance of 83.5\% and 46.69\% mIoU on Cityscapes and ADE20K, respectively.