Search

The semi-leptonic $b \to c l \nu$ processes are receiving a lot of attention, as the lepton flavor universality violation has been hinted by the measured ratios $R_{D^{(*)}} = \Gamma(B \to D^{(*)} \tau\nu)/\Gamma(B \to D^{(*)} \ell\nu)$ for $\ell = e,\mu$. Recently, it has also been pointed out that the baryonic counterpart, $R_{\Lambda_c} = \Gamma (\Lambda_b \to \Lambda_c \tau\nu)/\Gamma (\Lambda_b \to \Lambda_c \ell\nu)$, has a strong correlation with $R_{D^{(*)}}$, referred to as the R ratio sum rule in this paper. The correlation is almost independent of the new physics (NP) contributions and hence can predict $R_{\Lambda_c}$ from the measured $R_{D^{(*)}}$. On the other hand, we have fewer measurements and/or theoretical studies of the semi-leptonic $b \to u l \nu$ processes, although the same arguments can be applied to the ratios $R_\pi$, $R_\rho$, and $R_p$ as above. Since these processes are measurable at the ongoing LHCb run-3 and/or Belle~II experiments, precise studies on them are important as well. In this paper, we obtain the semi-analytic formulae for all the aforementioned $R_X$ ratios in the presence of model-independent NP contributions by using the available lattice QCD and/or light-cone sum rule fits to the form factors. Two novel points are highlighted: (i) We evaluate uncertainties of $R_X$ including both the Standard Model (SM) and NP terms, inherited from the form factor fits, and discuss how the uncertainties affect the $R$ ratio sum rules. (ii) We obtain the R ratio sum rule among the semi-leptonic $b \to u l \nu$ processes for the first time, which provides a complementary motivation for observing these processes. In addition, based on our model-independent results, we investigate how the different NP scenarios work in the $b\to c$ and $b\to u$ sectors and perform a combined study in the framework of SM effective field theory with specific flavor symmetries., Comment: 39 pages (main text), 4 figures, 6 tables

Reinforcement Learning (RL) empowers agents to acquire various skills by learning from reward signals. Unfortunately, designing high-quality instance-level rewards often demands significant effort. An emerging alternative, RL with delayed reward, focuses on learning from rewards presented periodically, which can be obtained from human evaluators assessing the agent's performance over sequences of behaviors. However, traditional methods in this domain assume the existence of underlying Markovian rewards and that the observed delayed reward is simply the sum of instance-level rewards, both of which often do not align well with real-world scenarios. In this paper, we introduce the problem of RL from Composite Delayed Reward (RLCoDe), which generalizes traditional RL from delayed rewards by eliminating the strong assumption. We suggest that the delayed reward may arise from a more complex structure reflecting the overall contribution of the sequence. To address this problem, we present a framework for modeling composite delayed rewards, using a weighted sum of non-Markovian components to capture the different contributions of individual steps. Building on this framework, we propose Composite Delayed Reward Transformer (CoDeTr), which incorporates a specialized in-sequence attention mechanism to effectively model these contributions. We conduct experiments on challenging locomotion tasks where the agent receives delayed rewards computed from composite functions of observable step rewards. The experimental results indicate that CoDeTr consistently outperforms baseline methods across evaluated metrics. Additionally, we demonstrate that it effectively identifies the most significant time steps within the sequence and accurately predicts rewards that closely reflect the environment feedback.

Toxicity classification in textual content remains a significant problem. Data with labels from a single annotator fall short of capturing the diversity of human perspectives. Therefore, there is a growing need to incorporate crowdsourced annotations for training an effective toxicity classifier. Additionally, the standard approach to training a classifier using empirical risk minimization (ERM) may fail to address the potential shifts between the training set and testing set due to exploiting spurious correlations. This work introduces a novel bi-level optimization framework that integrates crowdsourced annotations with the soft-labeling technique and optimizes the soft-label weights by Group Distributionally Robust Optimization (GroupDRO) to enhance the robustness against out-of-distribution (OOD) risk. We theoretically prove the convergence of our bi-level optimization algorithm. Experimental results demonstrate that our approach outperforms existing baseline methods in terms of both average and worst-group accuracy, confirming its effectiveness in leveraging crowdsourced annotations to achieve more effective and robust toxicity classification., Comment: 38th Conference on Neural Information Processing Systems (NeurIPS 2024)

We study the CP asymmetries of the rare top-quark decays $t \to c \gamma$ and $t \to cg$ in the aligned two-Higgs-doublet model (A2HDM), which is generically characterized by new sources of CP violation beyond the Standard Model (SM). Specifically, the branching ratios and CP asymmetries of these rare top-quark decays are explicitly formulated, with an emphasis on the origins of weak and strong phases in the A2HDM. Taking into account the most relevant constraints on this model, we evaluate the variations of these observables with respect to the model parameters. It is found that the branching ratios of $t \to c \gamma$ and $t \to cg$ decays can maximally reach up to $1.47\times10^{-10}$ and $4.86\times10^{-9}$ respectively, which are about four and three orders of magnitude higher than the corresponding SM predictions. While the branching ratios are almost independent of the relative phase $\varphi$ between the two alignment parameters $\varsigma_u$ and $\varsigma_d$ within the allowed parameter space, the CP asymmetries are found to be very sensitive to $\varphi$. When the two alignment parameters are complex with a non-zero $\varphi$ varied within the range $[50^\circ,150^\circ]$, the magnitudes of the CP asymmetries can be significantly enhanced relative to both the SM and the real case. In particular, the maximum absolute values of the CP asymmetries can even reach up to $\mathcal{O}(1)$ for these two decay modes, in the range $\varphi \in [70^\circ,100^\circ]$. These interesting observations could be utilized to discriminate the SM and the different scenarios of the A2HDM., Comment: 46 pages, 12 figures, and 3 tables

We present here the CP asymmetries in the decay rate and angular distributions of $\tau\to K_S\pi\nu_\tau$ decays in the Standard Model (SM) and beyond (BSM). The CP asymmetries in the SM are induced by the CP violation in $K^0-\bar{K}^0$ mixing. To investigate the BSM CP-violating (CPV) effects, a model-independent analysis is performed by using the low-energy effective field theory (LEFT) framework at $\mu=2$~GeV. If one further assumes the BSM physics to stem from above the electroweak scale, the LEFT shall then be matched onto the SM effective field theory (SMEFT), the operators of which contributing to $\tau\to K_S\pi\nu_\tau$ decays will also contribute to the neutron electric dipole moment (EDM) and $D^0-\bar{D}^0$ mixing. The stringent bounds from the latter suggest that no remarkable CPV effects can be observed in either the decay rate or the angular distributions. The prospects for future measurements of these observables are also mentioned., Comment: 8 pages, 3 figures, and 1 table; proceedings to the 2024 International Workshop on Future Tau Charm Facilities (FTCF2024)

In this work, we make a study of the $\tau^- \to \omega\pi^-\nu_\tau$ decay in the framework of low-energy effective field theory. The $J^{\mathcal{P}G}$ decompositions of the quark currents and the $\omega\pi$ final state show that, besides the Standard Model vector interaction, only the non-standard tensor interaction can have a non-zero contribution to the decay. To discuss its effect, a reliable calculation of the $\omega\pi$ tensor form factors is necessary. After constructing the Lagrangian of resonance chiral theory with external tensor sources, we calculate both the vector and tensor form factors with the relevant resonance couplings determined by combining the QCD short-distance constraints, the fit to the spectral function of $\tau^- \to \omega\pi^-\nu_\tau$ decay, as well as the matching between the $\mathcal{O}(p^4)$ odd-intrinsic-parity operators after integrating out the vector resonances and the $\mathcal{O}(p^6)$ operators of chiral perturbation theory. The new physics effect is then investigated in the distributions of the spectral function and the forward-backward asymmetry of $\tau^- \to \omega\pi^-\nu_\tau$ decay. We find that the spectral function is dominated by the Standard Model, and the non-standard tensor contribution is negligible. However, since the forward-backward asymmetry can be only generated with a non-zero tensor interaction, the observable is quite sensitive to this kind of new physics. A future measurement of the observable at the Belle II experiment as well as at the proposed Tera-Z and STCF facilities is, therefore, strongly called for to check the existence of such a non-standard tensor interaction., Comment: 27 pages, 4 tables, and 2 figures; minor modification, final version published in the journal

In this paper, we investigate an offline reinforcement learning (RL) problem where datasets are collected from two domains. In this scenario, having datasets with domain labels facilitates efficient policy training. However, in practice, the task of assigning domain labels can be resource-intensive or infeasible at a large scale, leading to a prevalence of domain-unlabeled data. To formalize this challenge, we introduce a novel offline RL problem setting named Positive-Unlabeled Offline RL (PUORL), which incorporates domain-unlabeled data. To address PUORL, we develop an offline RL algorithm utilizing positive-unlabeled learning to predict the domain labels of domain-unlabeled data, enabling the integration of this data into policy training. Our experiments show the effectiveness of our method in accurately identifying domains and learning policies that outperform baselines in the PUORL setting, highlighting its capability to leverage domain-unlabeled data effectively.

Action recognition, an essential component of computer vision, plays a pivotal role in multiple applications. Despite significant improvements brought by Convolutional Neural Networks (CNNs), these models suffer performance declines when trained with discontinuous video frames, which is a frequent scenario in real-world settings. This decline primarily results from the loss of temporal continuity, which is crucial for understanding the semantics of human actions. To overcome this issue, we introduce the 4A (Action Animation-based Augmentation Approach) pipeline, which employs a series of sophisticated techniques: starting with 2D human pose estimation from RGB videos, followed by Quaternion-based Graph Convolution Network for joint orientation and trajectory prediction, and Dynamic Skeletal Interpolation for creating smoother, diversified actions using game engine technology. This innovative approach generates realistic animations in varied game environments, viewed from multiple viewpoints. In this way, our method effectively bridges the domain gap between virtual and real-world data. In experimental evaluations, the 4A pipeline achieves comparable or even superior performance to traditional training approaches using real-world data, while requiring only 10% of the original data volume. Additionally, our approach demonstrates enhanced performance on In-the-wild videos, marking a significant advancement in the field of action recognition., Comment: Accepted by ECAI24, model and code are available at https://github.com/xingyu-song/4A

Recently, the Belle II collaboration announced the first measurement of $\mathcal B(B^+\to K^+\nu\bar\nu)$, which is found to be about $2.7\sigma$ higher than the SM prediction. We decipher the data with two new physics scenarios: the underlying $b\to s \nu\bar\nu$ transition is, besides the SM contribution, further affected by heavy new mediators that are much heavier than the EW scale, or amended by an additional decay channel with undetected light final states like dark matter or axion-like particles. These two scenarios can be most conveniently analyzed in the SMEFT and the dark SMEFT (DSMEFT) framework, respectively. We consider the flavour structures of the resulting effective operators to be either generic or satisfy the minimal flavour violation (MFV) hypothesis, both for the quark and lepton sectors. In the first scenario, once the MFV is assumed, only one SM-like low-energy effective operator induced by the SMEFT dim-6 operators can account for the Belle II excess, the parameter space of which is, however, excluded by the Belle upper bound on $\mathcal B(B^0\to K^{*0}\nu\bar\nu)$. In the second scenario, it is found that the Belle II excess can be accommodated by 22 of the DSMEFT operators involving one or two scalar, fermionic, or vector dark matters as well as ALPs. These operators also receive dominant constraints from the $B^0\to K^{*0}+$inv and $B_s\to$inv decays. In the MFV hypothesis, the number of viable operators is reduced to 14, and the $B^+\to\pi^+ +$inv and $K^+\to\pi^++$inv decays start to put further constraints. Within the parameter space allowed by all the current experimental data, the $q^2$ distributions (and $F_L$) of the $B\to K^{(*)}+$inv decays are studied for each viable operator. In addition, we, for the first time, calculate systematically the longitudinal polarization fraction $F_L$ of $K^*$ in the $B\to K^*+$inv decays within the DLEFT., Comment: 51 pages, 13 figures, comments welcome; v2: discussions on $F_L$ of the $B \to K^* + \text{inv}$ decay in the DSMEFT added, errors in the $B \to V + \text{inv}$ decays for some operators involving two scalar or vector DM fields fixed, the related discussions and figures 2, 6, 7, 11 and 12 corrected, main conclusion unchanged, a few comments and refs added

In Reinforcement Learning (RL), it is commonly assumed that an immediate reward signal is generated for each action taken by the agent, helping the agent maximize cumulative rewards to obtain the optimal policy. However, in many real-world scenarios, designing immediate reward signals is difficult; instead, agents receive a single reward that is contingent upon a partial sequence or a complete trajectory. In this work, we define this challenging problem as RL from Bagged Reward (RLBR), where sequences of data are treated as bags with non-Markovian bagged rewards, leading to the formulation of Bagged Reward Markov Decision Processes (BRMDPs). Theoretically, we demonstrate that RLBR can be addressed by solving a standard MDP with properly redistributed bagged rewards allocated to each instance within a bag. Empirically, we find that reward redistribution becomes more challenging as the bag length increases, due to reduced informational granularity. Existing reward redistribution methods are insufficient to address these challenges. Therefore, we propose a novel reward redistribution method equipped with a bidirectional attention mechanism, enabling the accurate interpretation of contextual nuances and temporal dependencies within each bag. We experimentally demonstrate that our proposed method consistently outperforms existing approaches.

Learning multi-lingual sentence embeddings is a fundamental task in natural language processing. Recent trends in learning both mono-lingual and multi-lingual sentence embeddings are mainly based on contrastive learning (CL) among an anchor, one positive, and multiple negative instances. In this work, we argue that leveraging multiple positives should be considered for multi-lingual sentence embeddings because (1) positives in a diverse set of languages can benefit cross-lingual learning, and (2) transitive similarity across multiple positives can provide reliable structural information for learning. In order to investigate the impact of multiple positives in CL, we propose a novel approach, named MPCL, to effectively utilize multiple positive instances to improve the learning of multi-lingual sentence embeddings. Experimental results on various backbone models and downstream tasks demonstrate that MPCL leads to better retrieval, semantic similarity, and classification performances compared to conventional CL. We also observe that in unseen languages, sentence embedding models trained on multiple positives show better cross-lingual transfer performance than models trained on a single positive instance., Comment: Accepted to EACL 2024, main conference

The absence of semitauonic decays of charmed hadrons makes the decay processes mediated by the quark-level $c\to d \tau^+ \nu_{\tau}$ transition inadequate for probing a generic new physics (NP) with all kinds of Dirac structures. To fill in this gap, we consider in this paper the quasielastic neutrino scattering process $\nu_{\tau}+n\to \tau^-+\Lambda_c$, and propose searching for NP through the polarizations of the $\tau$ lepton and the $\Lambda_c$ baryon. In the framework of a general low-energy effective Lagrangian, we perform a comprehensive analysis of the (differential) cross sections and polarization vectors of the process both within the Standard Model and in various NP scenarios, and scrutinize possible NP signals. We also explore the influence on our findings due to the uncertainties and the different parametrizations of the $\Lambda_c \to N$ transition form factors, and show that they have become one of the major challenges to further constrain possible NP through the quasielastic scattering process., Comment: 31 pages, 17 figures, and 3 tables. Comments are welcome

Motivated by the latest CDF $W$-mass measurement as well as the muon $g-2$ anomaly and the discrepancies observed in $b \to s \ell^+ \ell^-$ transitions, we propose an extension of the Standard Model (SM) with the $SU(2)_L$-singlet vector-like fermion partners that are featured by additional $U(1)^\prime$ gauge symmetry. The fermion partners have the same SM quantum numbers as of the right-handed SM fermions, and can therefore mix with the latter after the electroweak and the $U(1)^\prime$ symmetry breaking. As a result, desirable loop-level corrections to the $(g-2)_\mu$, the $W$-boson mass $m_W$ and the Wilson coefficient $C_9$ in $b \to s \mu^+ \mu^-$ transitions can be obtained. The final allowed parameter space is also consistent with the constraints from the $Z \to \mu^+ \mu^-$ decay, the neutrino trident production and the LHC direct searches for the vector-like quarks and leptons., Comment: 34 pages, 15 figures, comments welcome

We study the $B^- \to K^- \eta \eta_c$ decay by taking into account the $S$-wave contributions from the pseudoscalar meson-pseudoscalar meson interactions within the unitary coupled-channel approach, where the $D\bar{D}$ bound state is dynamically generated. In addition, the contribution from the intermediate resonance $K_0^*(1430)^-$, with $K_0^*(1430)^- \to K^-\eta$, is also considered. Our results show that there is a clear peak around $3720$~MeV in the $\eta \eta_c$ invariant mass distribution, which could be associated with the $D \bar{D}$ bound state. The future precise measurements of the $B^- \to K^- \eta \eta_c$ process at the Belle II and LHCb experiments could be, therefore, used to check the existence of the $D \bar{D}$ bound state, and to deepen our understanding of the hadron-hadron interactions., Comment: 14 pages, 9 figures

The semitauonic $B_c^- \to J/\psi \tau^-\bar{\nu}_\tau$ decay is optimal to scrutinize possible new physics effects in $b \to c \tau^- \bar{\nu}_\tau$ transitions as indicated by the current data on $R(D^{(*)})$ anomalies. In this work, we study the $B_c^- \to J/\psi \tau^-\bar{\nu}_\tau$ decay with both polarized $\tau$ and $J/\psi$. Their subsequent decays, with $J/\psi\to \mu^+ \mu^-$ as well as $\tau^- \to \pi^- \nu_\tau$, $\tau^- \to \rho^- \nu_\tau$ and $\tau^- \to \ell^-\bar{\nu}_\ell\nu_\tau$, are exploited to extract the energy and angular distributions of the charged final-state particles in the processes. Starting with the most general effective Hamiltonian relevant for the $b \to c \tau^- \bar{\nu}_\tau$ transitions, including all possible Lorentz structures of the dimension-six operators with both left- and right-handed neutrinos, we first derive the five-fold differential decay rate in terms of the visible final-state kinematics. From this distribution, we then construct in total 34 normalized observables, among which nine refer to the CP-violating triple product asymmetries that vanish within the Standard Model. We also construct five new observables based on the combinations of these normalized observables that can only be attributed to the right-handed neutrinos. On the other hand, considering the low statistics of the fully differential distribution, we introduce some integrated observables with only one kinematic variable left, which are more promising to be measured due to the largely increased statistics. The sensitivities of all these observables to the different new physics scenarios are investigated in detail. Finally, assuming an ideal circumstance, we give an estimate of the statistical uncertainties of the nine CP-conserving observables at LHCb and found that $\tau^-\to \pi^-\nu_\tau$ has the highest analyzing power among the three $\tau$ decay channels., Comment: 40 pages, 7 figures and 2 tables; SM predictions in different q2 bins added, final version published in the journal

The recently updated measurement of the $W$-boson mass by the CDF collaboration exhibits a $7\sigma$ deviation from the SM expectation, which may imply a sign of new physics beyond the SM. The observed discrepancy could be explained by new fermions that carry the electroweak gauge charges and affect the vacuum polarization of gauge bosons. Notably, if the new fermions also have the same quantum numbers as of the SM quarks, they can mix with the latter and thus modify the penguin diagrams governing the $b \to s \ell^+ \ell^-$ transitions. Therefore, the $W$-boson mass shift could be related to the $b \to s \ell^+ \ell^-$ anomalies observed by the LHCb collaboration during the past few years. To investigate this possibility, we consider in this paper a model containing a vector-like top partner gauged under a new $U(1)^\prime$ symmetry. It is found that the latest CDF $m_W$ measurement and the $b \to s \ell^+ \ell^-$ anomalies can be simultaneously accommodated at $2\sigma$ level., Comment: 7 pages, 3 figures, typos fixed, figures refined, and references added

If the leptoquarks proposed to account for the intriguing anomalies observed in the $B$-meson decays, $R_{D^{(\ast)}}$ and $R_{K^{(\ast)}}$, as well as in the anomalous magnetic moment of the muon, $(g-2)_\mu$, can be embedded into the scotogenic Dirac neutrino model, all these flavor anomalies, together with the origin of neutrino masses and the nature of dark matter, would be potentially addressed in a unified picture. Among the minimal seesaw, one-loop, and two-loop realizations of the dimension-4 effective operator $\mathcal{L}_{4}$ for the Dirac neutrino masses, we show that plenty of diagrams associated with the two-loop realizations of $\mathcal{L}_{4}$ can support the coexistence of leptoquarks and dark matter candidates. After a simple match of these leptoquarks to those that can accommodate all the flavor anomalies, we establish the scotogenic Dirac neutrino models embedded with leptoquarks, which could address all the problems mentioned above., Comment: 15 pages, 16 tables, 9 figures

We investigate the potential for searching and identifying the leptoquark (LQ) effects in the charm sector through the low-energy polarized scattering processes $\vec{e}^{\,-}p\to e^-\Lambda_c$, $e^-\vec{p}\to e^-\Lambda_c$, and $\vec{e}^{\,-}\vec{p}\to e^-\Lambda_c$. Considering only the longitudinally polarized processes, we show that the different LQ models can be disentangled from each other by measuring the four spin asymmetries, $A_{L}^e$, $A_{L}^p$, $A_{L3}^{ep}$, and $A_{L6}^{ep}$, constructed in terms of the polarized cross sections. Although it is challenging to accomplish the same goal with transversely polarized processes, we find that investing them in future experiments is especially beneficial, since they can directly probe into the imaginary part of the Wilson coefficients in the general low-energy effective Lagrangian. With our properly designed experimental setups, it is also demonstrated that promising event rates can be expected for all these processes and, even in the worst-case scenario -- no LQ signals are observed at all, they can still provide a competitive potential for constraining the new physics, compared with those from the conventional charmed-hadron weak decays and the high-$p_T$ dilepton invariant mass tails at high-energy colliders., Comment: 24 pages, 11 tables, and 7 figures; more discussions and references added, final version published in the journal

If the Dirac neutrino masses are generated by a new scalar doublet with a vacuum at keV order, there would be rare hope to probe the framework in low-energy flavor physics, such as the lepton-flavor violating processes. However, the predicted neutrino Yukawa couplings being around the order of electronic Yukawa can realize a purely thermal Dirac leptogenesis, and thus render a more direct link between the high- and the low-energy CP violation. Despite its inert property in the low-energy flavor processes, the keV-vacuum induced Dirac neutrino model can generate a significant shift of the effective neutrino number, which can in turn be probed by the big-bang nucleosynthesis and cosmic microwave background epochs. We show that such a keV-vacuum induced Dirac neutrino model is already constrained by the current observations and will be probed with forecast sensitivity, serving therefore as a complementary avenue to test the Dirac leptogenesis., Comment: 8 pages, and 3 figures

We update the Standard Model (SM) predictions for the branching ratios of the rare top-quark decays $t \to cg(g)$, and evaluate the maximum rates that can be reached in the aligned as well as in the four conventional two-Higgs-doublet models (2HDMs) with $\mathcal{Z}_2$ symmetries. Taking into account the relevant constraints on the model parameters resulting from a global fit obtained at the $95.5\%$ confidence level, we find that the branching ratios of $t \to cg$ and $t \to cgg$ decays can reach up to $3.36\times 10^{-9}$ and $2.95\times 10^{-9}$ respectively, being therefore of the same order, in the aligned 2HDM (A2HDM). This is obviously different from the SM case, where the predicted branching ratio of the three-body decay $t \to cgg$ is about two orders of magnitude larger than that of the two-body decay $t \to cg$. On the other hand, compared with the SM predictions, no significant enhancements are observed in the four conventional 2HDMs with $\mathcal{Z}_2$ symmetries for the branching ratios of these two decays. Nevertheless, the predicted branching ratios of $t \to cg$ and $t \to cgg$ decays in the A2HDM are still out of the expected sensitivities of the future high-luminosity Large Hadron Collider and the Future Circular Collider in hadron-hadron mode., Comment: 48 pages, 6 figures, and 4 tables; more references and discussions added, final version published in the journal