AI Sec Watch

TensorFlow, an open source platform for machine learning, has a vulnerability in its `tf.raw_ops.SparseDenseCwiseDiv` function where division by zero is not properly handled, causing the program to crash or behave unexpectedly. The vulnerability affects multiple older versions of TensorFlow that are still being supported.

TensorFlow versions up to 2.5.0 have a vulnerability where attackers can overwrite arbitrary files by providing a specially crafted archive when the tf.keras.utils.get_file function is used with the extract=True setting. This happens because the function doesn't properly validate file paths during extraction (a weakness called path traversal, where attackers manipulate file paths to access files outside intended directories). The vendor notes that this function was not designed to handle untrusted archives.

TensorFlow (an open-source platform for machine learning) has a bug where passing invalid arguments to a specific function called `tf.raw_ops.SparseCountSparseOutput` causes a segfault (a crash where the program tries to access memory it shouldn't). This happens because the function doesn't properly handle exceptional conditions (unexpected or invalid inputs).

TensorFlow (an open source machine learning platform) crashes when you pass a complex argument to the `tf.transpose` function while also using the `conjugate=True` argument. This happens because the software doesn't properly handle this unusual combination of inputs.

TensorFlow is a machine learning platform that had a vulnerability where an attacker could crash the system by sending invalid arguments to the `tf.strings.substr` function, which performs string operations. This vulnerability was caused by improper error handling (not properly catching and managing exceptional conditions that shouldn't happen).

TensorFlow, a machine learning platform, has a vulnerability where TrySimplify (a code optimization component) can crash by dereferencing a null pointer (trying to access memory that doesn't exist) when optimizing nodes with no inputs. This undefined behavior can cause the program to fail unexpectedly.

A vulnerability in TensorFlow (an open source machine learning platform) allows attackers to cause a stack overflow (a crash caused by a program using too much memory on the call stack) by sending specially crafted input to the `ParseAttrValue` function through recursion (when a function calls itself repeatedly).

A bug in TensorFlow's `tf.io.decode_raw` function causes incorrect results and crashes when using certain combinations of parameters. The problem stems from incorrect pointer arithmetic (moving through memory incorrectly), which causes the function to skip parts of input data and write outside the allocated memory bounds (OOB write, where data is written to memory locations it shouldn't access), potentially leading to crashes or more serious attacks.

TensorFlow (an open-source machine learning platform) has a vulnerability in its `tf.raw_ops.CTCLoss` function where incomplete validation (insufficient checking of input data) allows an attacker to trigger an OOB read from heap (accessing memory outside the intended boundaries). This is a memory safety issue that could crash the program or expose sensitive data.