Search

PurposeThis research uses rational choice theory to analyze the effects of motivation, premeditation and offender characteristics on offenders' weapons during decision-making processes when they are violent towards on-duty police officers. The paper aims to discuss the aforementioned issues.Design/methodology/approachThe researchers examined 597 cases (n = 597) of violence against the police in the China Judgments Online (CJO) database, and analyzed the data using multinomial logistic regression methods. Rational choice theory was used to explore the offenders' weapons decision-making process.FindingsThe research results showed that offenders with premeditation were more likely to use a weapon, and tended to choose sharp weapons; offenders motivated to “escape arrest” were more likely to use a weapon, and tended to choose a vehicle as a weapon; and offenders motivated by “conflict resolution” were more likely to choose a sharp or blunt weapon.Research limitations/implicationsThese findings have limited applicability to other countries and must be considered in the local background of violence against police.Practical implicationsThrough the rational choice theory analytical framework, this study clarifies how motivation and premeditation influence offenders' weapons decision-making processes.Social implicationsAlso, this study may provide support for frontline police officers' law enforcement.Originality/valueThe research identified some specific connections between offenders' weapon choice preferences, their motivation for the violence and whether or not there was premeditation. The findings provide guidance for police agencies developing preventive policies, and for frontline officers in interpreting and managing the situations they face.

PurposeBased on the theoretical viewpoints of criminal geography and environmental criminology, this research uses spatial multi-criteria decision-making methods. In the process of spatial decision-making and optimization of police resources, researchers fully consider the dynamic application of Geographic Information System (GIS) and the effects of spatial prevention and control.Design/methodology/approachResearchers use an integrated method combining Policing Geographic Information System (PGIS) and multi-criteria decision analysis (MCDA). On the one hand, police GIS has an excellent visual data analysis platform and integrated decision support system in data management, spatial analysis, data exploration and regression analysis. On the other hand, through the design of the indicator system, the quantification of indicators, the determination of weights, comprehensive evaluation and sensitivity analysis, MCDA can select the best plan from a large number of alternatives. When joining MCDA, the spatial dimension will bring the research results closer to the real world.FindingsThe study finds that the crime of burglary is affected to a certain extent by the distribution of police forces, the location of police units. Another important finding of this research is the correlation between more precise preventive measures and the crime of burglary.Originality/valueFrom a practical point of view, this research would help advance the role of police units and law enforcement agencies in preventing burglary crimes and provide experience for the allocation of regional police resources.

We study the problem of safe offline reinforcement learning (RL), the goal is to learn a policy that maximizes long-term reward while satisfying safety constraints given only offline data, without further interaction with the environment. This problem is more appealing for real world RL applications, in which data collection is costly or dangerous. Enforcing constraint satisfaction is non-trivial, especially in offline settings, as there is a potential large discrepancy between the policy distribution and the data distribution, causing errors in estimating the value of safety constraints. We show that na\"ive approaches that combine techniques from safe RL and offline RL can only learn sub-optimal solutions. We thus develop a simple yet effective algorithm, Constraints Penalized Q-Learning (CPQ), to solve the problem. Our method admits the use of data generated by mixed behavior policies. We present a theoretical analysis and demonstrate empirically that our approach can learn robustly across a variety of benchmark control tasks, outperforming several baselines., Comment: AAAI 2022; Spotlight @ RL4RealLife Workshop ICML2021

Attention and working memory (WM) are distinct cognitive functions, yet given their close interactions, it has been proposed that they share the same neuronal mechanisms. Here we show that in macaques performing a WM-guided feature attention task, the activity of most neurons in areas middle temporal (MT), medial superior temporal (MST), lateral intraparietal (LIP), and posterior lateral prefrontal cortex (LPFC-p) displays either WM coding or attentional modulation, but not both. One area thought to play a role in both functions is LPFC-p. To test this, we optogenetically inactivated LPFC-p bilaterally during the attention or WM periods of the task. Attention period inactivation reduced attentional modulation in LPFC-p, MST, and LIP neurons, and impaired task performance. WM period inactivation did not affect attentional modulation nor performance, and minimally reduced WM coding. Our results suggest that feature attention and WM have dissociable neuronal substrates, and that LPFC-p plays a critical role in attention but not WM.

Multilingual machine translation (MMT) benefits from cross-lingual transfer but is a challenging multitask optimization problem. This is partly because there is no clear framework to systematically learn language-specific parameters. Self-supervised learning (SSL) approaches that leverage large quantities of monolingual data (where parallel data is unavailable) have shown promise by improving translation performance as complementary tasks to the MMT task. However, jointly optimizing SSL and MMT tasks is even more challenging. In this work, we first investigate how to utilize intra-distillation to learn more *language-specific* parameters and then show the importance of these language-specific parameters. Next, we propose a novel but simple SSL task, concurrent denoising, that co-trains with the MMT task by concurrently denoising monolingual data on both the encoder and decoder. Finally, we apply intra-distillation to this co-training approach. Combining these two approaches significantly improves MMT performance, outperforming three state-of-the-art SSL methods by a large margin, e.g., 11.3\% and 3.7\% improvement on an 8-language and a 15-language benchmark compared with MASS, respectively, Comment: Findings of EACL 2023

Offline safe RL is of great practical relevance for deploying agents in real-world applications. However, acquiring constraint-satisfying policies from the fixed dataset is non-trivial for conventional approaches. Even worse, the learned constraints are stationary and may become invalid when the online safety requirement changes. In this paper, we present a novel offline safe RL approach referred to as SaFormer, which tackles the above issues via conditional sequence modeling. In contrast to existing sequence models, we propose cost-related tokens to restrict the action space and a posterior safety verification to enforce the constraint explicitly. Specifically, SaFormer performs a two-stage auto-regression conditioned by the maximum remaining cost to generate feasible candidates. It then filters out unsafe attempts and executes the optimal action with the highest expected return. Extensive experiments demonstrate the efficacy of SaFormer featuring (1) competitive returns with tightened constraint satisfaction; (2) adaptability to the in-range cost values of the offline data without retraining; (3) generalizability for constraints beyond the current dataset.

The prevalent use of benchmarks in current offline reinforcement learning (RL) research has led to a neglect of the imbalance of real-world dataset distributions in the development of models. The real-world offline RL dataset is often imbalanced over the state space due to the challenge of exploration or safety considerations. In this paper, we specify properties of imbalanced datasets in offline RL, where the state coverage follows a power law distribution characterized by skewed policies. Theoretically and empirically, we show that typically offline RL methods based on distributional constraints, such as conservative Q-learning (CQL), are ineffective in extracting policies under the imbalanced dataset. Inspired by natural intelligence, we propose a novel offline RL method that utilizes the augmentation of CQL with a retrieval process to recall past related experiences, effectively alleviating the challenges posed by imbalanced datasets. We evaluate our method on several tasks in the context of imbalanced datasets with varying levels of imbalance, utilizing the variant of D4RL. Empirical results demonstrate the superiority of our method over other baselines.

Most offline reinforcement learning (RL) methods suffer from the trade-off between improving the policy to surpass the behavior policy and constraining the policy to limit the deviation from the behavior policy as computing $Q$-values using out-of-distribution (OOD) actions will suffer from errors due to distributional shift. The recently proposed \textit{In-sample Learning} paradigm (i.e., IQL), which improves the policy by quantile regression using only data samples, shows great promise because it learns an optimal policy without querying the value function of any unseen actions. However, it remains unclear how this type of method handles the distributional shift in learning the value function. In this work, we make a key finding that the in-sample learning paradigm arises under the \textit{Implicit Value Regularization} (IVR) framework. This gives a deeper understanding of why the in-sample learning paradigm works, i.e., it applies implicit value regularization to the policy. Based on the IVR framework, we further propose two practical algorithms, Sparse $Q$-learning (SQL) and Exponential $Q$-learning (EQL), which adopt the same value regularization used in existing works, but in a complete in-sample manner. Compared with IQL, we find that our algorithms introduce sparsity in learning the value function, making them more robust in noisy data regimes. We also verify the effectiveness of SQL and EQL on D4RL benchmark datasets and show the benefits of in-sample learning by comparing them with CQL in small data regimes., Comment: ICLR 2023 notable top 5%

Additional file 1. Figure S1 Detection of transcript level of OsHAD3 in transgenic plants. Figure S2 Detection of transcript levels of other family members in WT and transgenic plants. Table S1 Primers used in this study.

Reward function is essential in reinforcement learning (RL), serving as the guiding signal to incentivize agents to solve given tasks, however, is also notoriously difficult to design. In many cases, only imperfect rewards are available, which inflicts substantial performance loss for RL agents. In this study, we propose a unified offline policy optimization approach, \textit{RGM (Reward Gap Minimization)}, which can smartly handle diverse types of imperfect rewards. RGM is formulated as a bi-level optimization problem: the upper layer optimizes a reward correction term that performs visitation distribution matching w.r.t. some expert data; the lower layer solves a pessimistic RL problem with the corrected rewards. By exploiting the duality of the lower layer, we derive a tractable algorithm that enables sampled-based learning without any online interactions. Comprehensive experiments demonstrate that RGM achieves superior performance to existing methods under diverse settings of imperfect rewards. Further, RGM can effectively correct wrong or inconsistent rewards against expert preference and retrieve useful information from biased rewards., Comment: Accept by ICLR2023. The first two authors contributed equally

Mixture-of-experts (MoE) models that employ sparse activation have demonstrated effectiveness in significantly increasing the number of parameters while maintaining low computational requirements per token. However, recent studies have established that MoE models are inherently parameter-inefficient as the improvement in performance diminishes with an increasing number of experts. We hypothesize this parameter inefficiency is a result of all experts having equal capacity, which may not adequately meet the varying complexity requirements of different tokens or tasks, e.g., in a multilingual setting, languages based on their resource levels might require different capacities. In light of this, we propose Stratified Mixture of Experts(SMoE) models, which feature a stratified structure and can assign dynamic capacity to different tokens. We demonstrate the effectiveness of SMoE on two multilingual machine translation benchmarks, where it outperforms multiple state-of-the-art MoE models. On a diverse 15-language dataset, SMoE improves the translation quality over vanilla MoE by +0.93 BLEU points on average. Additionally, SMoE is parameter-efficient, matching vanilla MoE performance with around 50\% fewer parameters.

Incorporating language-specific (LS) modules is a proven method to boost performance in multilingual machine translation. This approach bears similarity to Mixture-of-Experts (MoE) because it does not inflate FLOPs. However, the scalability of this approach to hundreds of languages (experts) tends to be unmanageable due to the prohibitive number of parameters introduced by full-rank matrices in fully-connected layers. In this work, we introduce the Language-Specific Matrix Synthesis (LMS) method. This approach constructs LS modules by generating low-rank matrices from two significantly smaller matrices to approximate the full-rank matrix. Furthermore, we condense multilingual knowledge from multiple LS modules into a single shared module with the Fuse Distillation (FD) technique to improve the efficiency of inference and model serialization. We show that our LMS method significantly outperforms previous LS methods and MoE methods with the same amount of extra parameters, e.g., 1.73 BLEU points over the Switch Transformer on many-to-many multilingual machine translation. Importantly, LMS is able to have comparable translation performance with much fewer parameters.

Interaction between active materials and the boundaries of geometrical confinement is key to many emergent phenomena in active systems. For living active matter consisting of animal cells or motile bacteria, the confinement boundary is often a deformable interface, and it has been unclear how activity-induced interface dynamics might lead to morphogenesis and pattern formation. Here we studied the evolution of bacterial active matter confined by a deformable boundary. We discovered that an ordered morphological pattern emerged at the interface characterized by periodically-spaced interfacial protrusions; behind the interfacial protrusions, bacterial swimmers self-organized into multicellular clusters displaying +1/2 nematic defects. Subsequently, a hierarchical sequence of transitions from interfacial protrusions to creeping branches allowed the bacterial active drop to rapidly invade surrounding space with a striking self-similar branch pattern. We found that this interface patterning is controlled by the local curvature of the interface, a phenomenon we denote as collective curvature sensing. Using a continuum active model, we revealed that the collective curvature sensing arises from enhanced active stresses near high-curvature regions, with the active length-scale setting the characteristic distance between the interfacial protrusions. Our findings reveal a protrusion-to-branch transition as a novel mode of active matter invasion and suggest a new strategy to engineer pattern formation of active materials.

The current state-of-the-art for few-shot cross-lingual transfer learning first trains on abundant labeled data in the source language and then fine-tunes with a few examples on the target language, termed target-adapting. Though this has been demonstrated to work on a variety of tasks, in this paper we show some deficiencies of this approach and propose a one-step mixed training method that trains on both source and target data with \textit{stochastic gradient surgery}, a novel gradient-level optimization. Unlike the previous studies that focus on one language at a time when target-adapting, we use one model to handle all target languages simultaneously to avoid excessively language-specific models. Moreover, we discuss the unreality of utilizing large target development sets for model selection in previous literature. We further show that our method is both development-free for target languages, and is also able to escape from overfitting issues. We conduct a large-scale experiment on 4 diverse NLP tasks across up to 48 languages. Our proposed method achieves state-of-the-art performance on all tasks and outperforms target-adapting by a large margin, especially for languages that are linguistically distant from the source language, e.g., 7.36% F1 absolute gain on average for the NER task, up to 17.60% on Punjabi., Accepted at Findings of NAACL 2022

The effects of S and Se treatment on cabbage, especially the interactions of S and Se metabolism with the biosynthesis of glucosinolate (GSL), including glucoraphanin, which is a major aliphatic GSL in cruciferous vegetables and the precursor of the anticancer compound sulforaphane, were examined. Cabbage plants were treated with sulfate and selenite (SeO32−), and the total S, Se, and GSL contents of cabbage head and outer foliage leaves were measured. Results showed that selenite treatment was beneficial to GSL biosynthesis and Se accumulation in cabbage head and outer foliage leaves. GSL synthesis was induced by exogenous selenite-elevated sulfate treatment at certain concentration ratios, i.e., 50-μΜ selenite + 1-mΜ sulfate or 100-μΜ selenite + 4-mΜ sulfate. A high exogenous sulfate concentration was more favorable to GSL accumulation than a low sulfate concentration. According to the relative expression of genes on GSL synthesis, an increase in the GSL content was attributed to the upregulation of gene expression and possible transportation from the outer foliage leaf to the head of cabbage. These results might be helpful for increasing the health benefits of cabbage by supplying exogenous S and Se. Further research should explore the effects of sulfate and selenite on GSL precursor substances to reveal the reason why total GSL contents increased.

Zero-shot cross-lingual information extraction (IE) describes the construction of an IE model for some target language, given existing annotations exclusively in some other language, typically English. While the advance of pretrained multilingual encoders suggests an easy optimism of "train on English, run on any language", we find through a thorough exploration and extension of techniques that a combination of approaches, both new and old, leads to better performance than any one cross-lingual strategy in particular. We explore techniques including data projection and self-training, and how different pretrained encoders impact them. We use English-to-Arabic IE as our initial example, demonstrating strong performance in this setting for event extraction, named entity recognition, part-of-speech tagging, and dependency parsing. We then apply data projection and self-training to three tasks across eight target languages. Because no single set of techniques performs the best across all tasks, we encourage practitioners to explore various configurations of the techniques described in this work when seeking to improve on zero-shot training., EMNLP 2021

Linear embedding transformation has been shown to be effective for zero-shot cross-lingual transfer tasks and achieve surprisingly promising results. However, cross-lingual embedding space mapping is usually studied in static word-level embeddings, where a space transformation is derived by aligning representations of translation pairs that are referred from dictionaries. We move further from this line and investigate a contextual embedding alignment approach which is sense-level and dictionary-free. To enhance the quality of the mapping, we also provide a deep view of properties of contextual embeddings, i.e., anisotropy problem and its solution. Experiments on zero-shot dependency parsing through the concept-shared space built by our embedding transformation substantially outperform state-of-the-art methods using multilingual embeddings.

Fine-tuning is known to improve NLP models by adapting an initial model trained on more plentiful but less domain-salient examples to data in a target domain. Such domain adaptation is typically done using one stage of fine-tuning. We demonstrate that gradually fine-tuning in a multi-stage process can yield substantial further gains and can be applied without modifying the model or learning objective., Comment: Adapt-NLP, EACL 2021

Additional file 1: Fig. S1. CLP induced AKI. Mice were subjected to CLP, and the kidney samples were collected at the indicated time. (A-B) Shown are representative Hematoxylin-eosin (HE) stains and Periodic Acid-Schiff (PAS) stain. (C) Histological analyses of renal tubular injury. (D) Quantitative analyses of IHC stain of NGAL. (E) Representative immunohistochemistry (IHC) images for NGAL. (F-G) Quantitative analyses of Scr and BUN. n = 6 mice/group, mean �� SD were presented. **P < 0.01.

Most prior approaches to offline reinforcement learning (RL) utilize \textit{behavior regularization}, typically augmenting existing off-policy actor critic algorithms with a penalty measuring divergence between the policy and the offline data. However, these approaches lack guaranteed performance improvement over the behavior policy. In this work, we start from the performance difference between the learned policy and the behavior policy, we derive a new policy learning objective that can be used in the offline setting, which corresponds to the advantage function value of the behavior policy, multiplying by a state-marginal density ratio. We propose a practical way to compute the density ratio and demonstrate its equivalence to a state-dependent behavior regularization. Unlike state-independent regularization used in prior approaches, this \textit{soft} regularization allows more freedom of policy deviation at high confidence states, leading to better performance and stability. We thus term our resulting algorithm Soft Behavior-regularized Actor Critic (SBAC). Our experimental results show that SBAC matches or outperforms the state-of-the-art on a set of continuous control locomotion and manipulation tasks.

Additional file 3: Fig. S3. MCTR1 inhibited erastin-induced ferroptosis by Nrf2 signaling. HK-2 cells were transfected with or without Nrf2 siRNA (30 nM) for 48 h and then treated with MCTR1 (100 nM) 0.5 h, followed by erastin (1 ��M) for 24 h. (A, C) Fold change of cell viability. (B, D) Visualization of cell viability was evaluated by phase-contrast microscopy. n = 3, mean �� SD were presented. **P < 0.01.

Additional file 2: Fig. S2. Specific inhibitor of ferroptosis ameliorates LPS-induced cell death. (A) Viability curves for HK-2 cells treated with different concentrations (0, 0.001, 0.01, 0.1, 1, 10 and 100 ug/ml) of LPS for 24 h. (B) Viability curves for HK-2 cells treated with LPS (1 ug/ml) for different times (0, 1.5, 3, 6, 12, and 24 h). HK-2 cells were pretreated with or without liproxstatin-1(10 ��M) or DFO (100 ��M) 0.5 h, followed by LPS (1 ug/ml) for 8 h. (C) Fold change of cell viability. (D) Visualizationof cell viability were evaluated by phase-contrast microscopy. n = 3, mean �� SD werepresented. **P < 0.01.

Both Raman and photoluminescence spectroscopies have been employed to investigate the variation of lattice vibrational and electronic properties of bilayer WSe2 for a large strain range. Upon application of uniaxial tensile strain, the initially degenerate in-plane E2g Raman mode evolves into two discrete modes as a consequence of inversion symmetry breaking. The decrease of direct bandgap energy accompanied by an increase of indirect bandgap energy leads to an indirect-to-direct bandgap crossover, which expectedly occurs at 2% strain. However, under higher strain, a phenomenon of strain relaxation appears in the WSe2 crystal. Such relaxation gives rise to a cease of further redshift of direct bandgap and a significant increase of direct excitonic linewidth. Our findings can provide a deeper understanding in the effects of strain engineering in WSe2 and are believed to be applicable to other layered transition-metal dichalcogenides.

In this study, the finite element model of the ultrasonic cleaning system was established and the harmonic response analysis was carried out to obtain the vibration distribution of the ultrasonic cleaning floor and the sound pressure distribution in the water. Scanning electron microscopy was used to observe the adhesion of fouling organisms. The results show that the ultrasonic treatment mode has a significant effect on the microscopic fouling adsorption of PMMA surface. When the total time of the ultrasonic action is same, the shorter the ultrasonic interval time, the more significant the effect of ultrasonic decontamination.

Modern out-of-order processors have increased capacity to exploit instruction level parallelism (ILP) and memory level parallelism (MLP), e.g., by using wide superscalar pipelines and vector execution units, as well as deep buffers for in-flight memory requests. These resources, however, often exhibit poor utilization rates on workloads with large working sets, e.g., in-memory databases, key-value stores, and graph analytics, as compilers and hardware struggle to expose ILP and MLP from the instruction stream automatically. In this paper, we introduce the IMLP (Instruction and Memory Level Parallelism) task programming model. IMLP tasks execute as coroutines that yield execution at annotated long-latency operations, e.g., memory accesses, divisions, or unpredictable branches. IMLP tasks are interleaved on a single thread, and integrate well with thread parallelism and vectorization. Our DSL embedded in C++, Cimple, allows exploration of task scheduling and transformations, such as buffering, vectorization, pipelining, and prefetching. We demonstrate state-of-the-art performance on core algorithms used in in-memory databases that operate on arrays, hash tables, trees, and skip lists. Cimple applications reach 2.5x throughput gains over hardware multithreading on a multi-core, and 6.4x single thread speedup., Comment: To appear in PACT'18

We examine the audit pricing of terrorism. We find that in terrorism-intense countries, auditors charge higher audit fees. A similar result is obtained using the September 11, 2001 terrorist attack in the US as a quasi-natural experiment. These findings suggest that auditors incorporate terrorism into risk assessment and charge a fee premium accordingly. Cross-sectional analyses reveal that shareholder litigation strengthens the impact of terrorism on audit pricing. However, such an impact weakens as client importance increases. Finally, we show that in terrorism-intense countries, auditors are more likely to issue non-clean opinions. We conclude that auditors price terrorism.

An analytical method was developed to extract and determine 2-methyl-4-chlorophenoxyacetic acid(MCPA) residue in sugarcane and soil using solid-phase extraction and high-performance liquid chromatography(SPE-HPLC).The MCPA residue in sugarcane and soil was dissolved in 0.04 mol/L sodium hydroxide aqueous solution,and purified through basic alumina-SPE column.Then the sample was detected by HPLC-UV.The method could be used to determine the MCPA residue in sugarcane and soil with fine linear relationship(R2=0.999 3) when the concentration range of MCPA was 0.05-10 mg/L and the limit of detection was 0.01 mg/kg.The recovery rates of MCPA from sugarcane and soil fortified with MCPA at 0.05,0.1,1.0 mg/kg were 82.2%-89.1% and 84.0%-85.7% respectively,with relative standard deviation10%.The results showed that the half-life of degradation for MCPA was 1.78-2.69 d in sugarcane,2.40-3.64 d in soil.

This paper reports the first high power tests of hybrid photonic band gap (PBG) accelerator structures. Three hybrid PBG (HPBG) structures were designed, built and tested at 17.14 GHz. Each structure had a triangular lattice array with 60 inner sapphire rods and 24 outer copper rods sandwiched between copper disks. The dielectric PBG band gap map allows the unique feature of overmoded operation in a TM₀₂ mode, with suppression of both lower order modes, such as the TM₁₁ mode, as well as higher order modes. The use of sapphire rods, which have negligible dielectric loss, required inclusion of the dielectric birefringence in the design. The three structures were designed to sequentially reduce the peak surface electric field. Simulations showed relatively high surface fields at the triple point as well as in any gaps between components in the clamped assembly. The third structure used sapphire rods with small pin extensions at each end and obtained the highest gradient of 19 MV/m, corresponding to a surface electric field of 78 MV/m, with a breakdown probability of 5 × 10⁻¹ per pulse per meter for a 100-ns input power pulse. Operation at a gradient above 20 MV/m led to runaway breakdowns with extensive light emission and eventual damage. For all three structures, multipactor light emission was observed at gradients well below the breakdown threshold. This research indicated that multipactor triggered at the triple point limited the operational gradient of the hybrid structure., United States. Department of Energy (Grant DE-SC0010075)