AI Sec Watch

TensorFlow, an open source machine learning platform, has a bug where invalid input to the `SparseMatrixNNZ` function (a function that counts non-zero values in a sparse matrix, which is a matrix stored efficiently by only keeping non-zero elements) causes the program to crash with a CHECK fail (an assertion error, where the program stops because a required condition wasn't met). This vulnerability affects multiple versions of TensorFlow.

TensorFlow (an open source machine learning platform) has a security vulnerability in its FractionalMaxPool and FractionalAvgPool functions when given invalid pooling_ratio values. Attackers can exploit this to access heap memory (the computer's temporary storage area outside normal program control), potentially causing the system to crash or allowing remote code execution (running harmful commands on someone else's computer).

TensorFlow (an open source machine learning platform) has a bug where certain inputs with incorrect dimensions crash the SdcaOptimizer component due to a failed validation check. This happens when `dense_features` or `example_state_data` inputs don't have the expected 2D structure (rank 2, meaning a table with rows and columns).

TensorFlow, an open source machine learning platform, crashes when a function called `SparseFillEmptyRowsGrad` receives empty inputs instead of data. This happens because the code doesn't properly validate (check) what data it receives before trying to process it.

TensorFlow (an open-source machine learning platform) crashes when a function called `FractionMaxPoolGrad` receives oversized inputs for `row_pooling_sequence` and `col_pooling_sequence` parameters. This is caused by an out-of-bounds read (accessing memory locations outside the intended range), which allows the program to fail unexpectedly.

TensorFlow (an open-source platform for machine learning) has a vulnerability where a function called `ThreadUnsafeUnigramCandidateSampler` crashes if it receives an input value for `filterbank_channel_count` that exceeds the maximum allowed size. This is caused by improper input validation (failure to check that user-provided values are within acceptable limits).

TensorFlow, an open source machine learning platform, has a vulnerability where the `MirrorPadGrad` function crashes with a heap OOB error (out-of-bounds memory access, where the software tries to read memory it shouldn't) when given incorrectly sized input padding values. This bug allows attackers to potentially crash TensorFlow applications.

TensorFlow Lite's `CONV_3D_TRANSPOSE` operator (a component that flips and reorganizes 3D data during machine learning processing) had a bug where it incorrectly calculated memory addresses when adding bias values, potentially allowing an attacker to write data outside the intended memory area (buffer overflow, where data gets written beyond allocated boundaries). The vulnerability only affects users who employ TensorFlow's default kernel resolver in their interpreter.

TensorFlow, an open source machine learning platform, has a vulnerability in the `tf.raw_ops.TensorListResize` function where providing a nonscalar value (a value that isn't a single number) for the `size` input causes a CHECK fail, which can be exploited to trigger a denial of service attack (making the system crash or become unavailable).