They are stated in the paragraph. Processes, devices, and network sockets (and others). These are objects managed by the kernel but appear as items in the file system and you can access them through specific paths. Their behavior also resemble regular files in many aspects. For example, you can read from and write to a device file (see below) just as you do with regular files.
Each process has its folder
$PPID is the process id. You can find the command line for the process in
/proc/$PPID/cmdline, for example.
Devices appear as device files. For example, if your hard disk is connected to a SCSI controller, you are likely to find it at
/dev/sda and the first partition as
/dev/sda1. So if you read from
/dev/sda you are actually reading the raw data on the disk. This can be useful when making an image.
/dev/tcp/$host/$port to identify TCP ports and alike for UDP. In general one can find information about sockets in
Standard input is also a file and you can find it in
/dev/stdin, similar for standard output. You can read from the file
/dev/zero to get zeros and from
/dev/random to get really high-quality random number (although very slowly). Pipe is frequently used in the shell and you can also create a named pipe. There are many more.
The answer above is based on Ubuntu LTS 16.04, other Linux/Unix versions may have different ways of indexing these objects.