I am a theoretician working primarily on astrophysics and general relativity. My research has touched many fields in high-energy astrophysics and cosmology, including gamma-ray bursts, supernovae, neutron star mergers, black hole physics, accretion process, gravitational lensing, creation of the universe, and time travel.

Richard Gott and I proposed a model for the self-creation of the Universe, in which the Universe started from an epoch with closed timelike curves so that time has no beginning. The model can naturally explain why there is a time arrow today. With Bohdan Paczynski, I proposed the first model for electromagnetic radiation from neutron star mergers. The predicted transient event (called minisupernovae, macronovae, or kilonovae) was observationally discovered in 2017 during the follow-up observation of the gravitational wave source GW170817. With Ramesh Narayan et al., I invented a KERRBB program for calculating the blackbody spectrum of a relativistic accretion disk around a Kerr black hole, which is now a standard tool for fitting the spectra of X-ray binaries and determining the spin parameter of black holes. Later, we extended the model to include calculation of the polarization of the blackbody radiation of a thin disk around a Kerr black hole.

Based on a sample of then available four pairs of gamma-ray bursts and supernovae with spectroscopically confirmed connection, I discovered a remarkable relation between the peak spectral energy of gamma-ray bursts and the peak bolometric luminosity of the underlying supernovae. The relation, which may have important implications for the nature of the gamma-ray burst and supernova connection, was later confirmed by the XRF080109 associated with a normal Type Ib SN2008D discovered in the spiral galaxy NGC2770. With a Swift sample of gamma-ray bursts with measured redshifts I studied the relation between gamma-ray bursts and star formation. I found that long-duration gamma-ray bursts trace both star formation and metallicity evolution in the Universe. The result indicates that if the star formation history can be accurately measured, it will be possible to determine the cosmic metallicity evolution with the redshift distribution of long-duration gamma-ray bursts.

To explain the existence of black holes of a billion (or even larger) solar masses at redshift six, I proposed that black holes in quasars have undergone a two-phase growing process: with a short super-Eddington accretion process they get their masses inflated by a very large factor until the feedback process becomes important, then with a prolonged sub-Eddington accretion process they have their masses increased by a factor of two or so. The overall average efficiency of this two-phase process is estimated to be ≳ 0.1, which is consistent with observational constraints.

Recently I became interested in electrodynamics in curved spacetime and unification of electromagnetic and gravitational forces. With some evidence I argued for an electromagnetic field equation with a term of the electromagnetic potential vector coupled to the spacetime curvature tensor. In particular, I showed that such a field equation can be derived from a five-dimensional Einstein field equation, if our four-dimensional spacetime is a hypersurface imbedded in a five-dimensional bulk spacetime.