Luminous and ultra-luminous infrared galaxies ((U)LIRGs) are rare today but are increasingly abundant at high redshifts. They are believed to be dusty starbursts, and hence should have high rates of supernovae (multiple events per year). Due to their extremely dusty environment, however, such supernovae could only be detected in restframe infrared and longer wavelengths, where our current facilities lack the capability of finding them individually beyond the local universe. We propose a new technique for higher redshifts, which is to search for the presence of supernovae through the variability of the integrated rest-frame infrared light of the entire hosts. We present a pilot study to assess the feasibility of this technique. We exploit a unique region, the "IRAC Dark Field" (IDF), that the Spitzer Space Telescope has observed for more than 14 years in 3--5 micron. The IDF also has deep far-infrared data (200--550 micron) from the Herschel Space Observatory that allow us to select high-redshift (U)LIRGs. We obtain a sample of (U)LIRGs that have secure optical counterparts, and examine their light curves in 3--5 micron. While the variabilities could also be caused by AGNs, we show that such contaminations can be identified. We present two cases where the distinct features in their light curves are consistent with multiple supernovae overlapping in time. Searching for supernovae this way will be relevant to the James Webb Space Telescope (JWST) to probe high-redshift (U)LIRGs into their nuclear regions where JWST will be limited by its resolution.

Brightest Cluster Galaxies (BCGs) are believed to have assembled most of their stars early in time and, therefore, should be passively evolving at low redshifts and appear "red-and-dead." However, there have been reports that a minority of low-redshift BCGs still have ongoing star formation rates (SFR) of a few to even $\sim$100 $M_\odot/yr$. Such BCGs are found in "cool-core" ("CC") clusters, and their star formation is thought to be fueled by "cooling flow." To further investigate the implications of low-redshift, star-forming BCGs, we perform a systematic search using the 22$\mu$m data ("W4" band) from the Wide-field Infrared Survey Explorer (WISE) on the GMBCG catalog, which contains 55,424 BCGs at $0.1\lesssim z\lesssim 0.55$ identified in the Sloan Digital Sky Survey (SDSS). Our sample consists of 389 BCGs that are bright in W4 ("W4BCGs"), most being brighter than 5 mJy. While some ($\lesssim 20\%$) might host AGN, most W4BCGs should owe their strong mid-IR emissions to dust-enshrouded star formation. Their median total IR luminosity ($L_{IR}$) is $5\times10^{11} L_{\odot}$ (SFR $\sim$50 $M_{\odot}/yr)$, and 27\% of the whole sample has $L_{IR}>10^{12} L_{\odot}$ (SFR $>$100 $M_{\odot}/yr$). Using ten W4BCGs that have Chandra X-ray data, we show that seven of them are possibly in CC clusters. However, in most cases (five out of seven) the mass deposition rate cannot account for the observed SFR. This casts doubt to the idea that cooling flows are the cause of the star formation in non-quiescent BCGs.