Abstract
The unsolved puzzles in nature suggest that the Standard Model (SM) is yet incomplete. We are now dedicating to validate the SM to a high precision and search for new physics (NP). In this thesis, we argue that colliders offer excellent experimental facilities with new opportunities emerging from the strong power of future colliders and advanced data analysis techniques. We focus on flavor physics and demonstrate that the Z factory run at future circular lepton colliders such as CEPC or FCC-ee provides an opportunity to test lepton flavor universality (LFU). We conduct a systematic study of LFU-violating observables and explore their sensitivity robustness against detector performance and their potential improvement with the message of event shape or beyond b-hadron decays. We interpret these results in a model-independent approach and constrain the SM effective field theory (SMEFT) Wilson coefficients to ≲ O(1) with the cut-off scale up to O(10)TeV. We also show new opportunities from the data analysis side by using novelty detection techniques to search for NP. Our approach exploits the complementarity between clustering-based and isolation-based algorithms, leading to significant performance improvements and greater applicability of novelty detection at colliders. We demonstrate our scheme with various gaussian samples and apply it to detect two significantly different signals at the LHC featuring a t¯tγγ final state: t¯th, giving a narrow resonance in the diphoton mass spectrum, and gravity-mediated supersymmetry, resulting in broad distributions at high transverse momentum. Compared to existing dedicated searches at the LHC, our scheme shows encouraging sensitivities for detecting both signals.